Nat Rev Immunol: cytokine release syndrome and associated neurotoxicity in tumor immunotherapy

summary

Paradigm shift has recently taken place in the field of oncology therapy, where cellular immunotherapy based on T cell engineering has added traditional anti-cancer drugs such as chemotherapy, radiotherapy, and small molecule drugs that target specific signaling pathways
.
Based on scientific developments in tumor immunology, genetic engineering, and cell production, novel patient-specific cell therapies have been rapidly adopted, as evidenced by the curative potential of chimeric antigen receptor (CAR)T cell therapies targeting CD19-expressing malignancies
.
However, the clinical benefit may be costly for many patients, with up to one third of patients experiencing significant toxicities directly associated with inducing potent immune effect responses, the most common of which are cytokine release syndromes and immune effector cell-associated neurotoxicity syndromes
.
This review discusses our current understanding of its pathophysiology and clinical features, as well as the development of novel therapeutic agents
for their prevention and/or management.

Adoptively metastatic tumor antigen-specific T cells are genetically engineered to express chimeric antigen receptors (CARS), resulting in durable clinical remission in patients with some types of tumors, particularly refractory and relapsed B-cell malignancies
expressing CD19.
CARs are synthetic antigen-recognizing receptors, including antibody-derived single-stranded variable fragments (scFv), hinge and transmembrane domains, and intracellular signaling domains
.
The intracellular signaling domain is generally CD3ζ signaling domain and co-stimulatory signaling domain, such as CD28 and 4-1BB (also known as TNFRSF9), and the intracellular signaling domain and the cellular phenotype of engineered T cells once activated by antigen-expressing target cells will affect the specific cytokine secretion and in vivo proliferation ability
of CAR T cells.

Despite their clinical success, CAR T cells also cause significant toxicity, which is directly related to
the robust immune effect response they induce.
Two major toxicities of CAR T cell production were not shown in early mouse models, but eventually occurred in clinical trials, namely cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS; Often referred to as neurotoxicity).

CRS usually begins with fever and constitutional symptoms such as chills, malaise, and anorexia, which can be high and persist for several days
.
In severe cases, CRS presents with other features of the systemic inflammatory response, including hypotension, hypoxia, and/or organ dysfunction, which may be secondary to hypotension or hypoxia, but can also result
from a direct effect of cytokine release.
Dysfunction of all major organ systems, including the heart, lung, liver, kidney, and gastrointestinal systems, can occur in patients with CRS, but organ dysfunction is preventable or reversible in most patients if the signs and symptoms of CRS are recognized and managed promptly
.

ICANS typically presents with toxic encephalopathy that begins with difficulty finding words, confusion, speech difficulties, aphasia, impaired fine motor skills, and lethargy, and seizures, motor weakness, cerebral oedema, and coma
may occur in severe cases.
Most patients with clinical features of ICANS have had CRS in the past, so CRS is considered a 'initiating event' or cofactor
of ICANS.
ICANS occurs more often after CRS symptoms resolve (Figure 1), and although less frequent, CRS and ICANS can occur simultaneously
.
In addition, like CRS, ICANS is reversible in most patients without permanent neurologic
deficits.

At present, most clinical applications of CAR T cells are for high-risk malignant diseases, and in the field of hemato-oncology, CAR T cells (CD19CAR T cells) targeting CD19 are mostly used for CAR T cells, so this is also the focus
of this review.
However, it should be noted that toxicity similar to that described here has been observed in multiple clinical indications for CAR T, with broad CAR T cell specificity
.
A better understanding of the clinical features and pathophysiology of CRS and ICANS is important as the development of genetically modified immune cells for non-malignant indications such
as autoimmune diseases and solid organ transplant tolerance is advancing by leaps and bounds, where the risk-benefit balance is not detrimental to potential toxicity, and the need to improve toxicity management or reduce toxicity of immunotherapy is important.

Here we discuss the known data to date and describe how it affects the development
of novel therapeutics to prevent and/or manage these toxicities.
In addition, the in-depth understanding of CRS and ICANS has broader implications for other systemic cytokine-mediated inflammatory diseases such as septic shock, macrophage activation syndrome (BOX 1), neuroinflammatory diseases, and emerging pathogen infections such as SARS-CoV-2 (BOX 2), as well as other therapeutic interventions such as IL-1 blockade in tumor immunotherapy (BOX 3).

。 As discussed below, our understanding of the pathophysiology of CRS and ICANS associated with CAR T-cell therapy began in the clinic and ended in laboratory studies and is now expanded
through exhaustive studies in animal models.

Pathophysiology of CRS

The pathophysiology of CRS can be divided into 5 main stages
.
In Phase 1, CAR T cells are infused into the patient and transferred to the tumor site, and CAR mediates recognition of antigen-expressing target cells
.
In stage 2, CAR T cells proliferate at the tumor site, and activated CAR T cells and cellular components of the tumor microenvironment can generate cytokines in situ, activate "bystander" endogenous immune cells, directly and indirectly kill tumor cells and develop CRS
.
In stage 3, elevated cytokines in peripheral blood and expansion of the CAR T cell population are associated with systemic inflammatory responses and can lead to endothelial damage and vascular leakage in multiple tissues and organs and their associated effects, including hypoxia, hypotension, and/or organ damage
.
In stage 4, cytokine diffusion and CAR T cell migration, endogenous T cells and peripherally activated monocytes enter the cerebrospinal fluid (CSF) and central nervous system (CNS), including disruption of the blood-brain barrier (BBB), consistent with
ICANS episodes.
In stage 5, activation-induced T cell death and tumor eradication result in decreased serum cytokines and a decreased systemic inflammatory response, end of CRS and/or ICANS symptoms, and possibly persistence of long-memory CAR T cells
.
The relative timing of onset and duration of CRS and ICANS is shown
in Figure 1.

CRS did not occur in preclinical studies of CD19CAR T cells and was only discovered
after initiating a phase I clinical trial of CAR T cells.
Although fever and hypotension in patients are associated with immediate elevation of serum cytokines, the underlying mechanism of this episodic outcome (more common in patients with high tumor burden) is initially unclear
.
It is now known that almost all patients receiving CD19CAR T-cell therapy develop some degree of CRS; In registered trials, up to one third of patients with B-cell acute lymphoblastic leukaemia developed severe CRS, and up to half developed ICANS, but the actual incidence of CRS and ICANS may be lower than this
.
Empirical testing of different blocking antibodies quickly identified IL-6 as a key mediator of CRS, making tocilizumab, a monoclonal antibody that blocks signaling through the IL-6 receptor (IL-6R), the mainstay
of CRS management.
Since activated T cells produce IL-6, it was initially assumed that CAR T cells themselves were the primary source, and although observations suggested additional possible contributing factors, subsequent studies identified macrophages and monocyte lineage cells as sources of
IL-6.
Notably, CRS associated with both CAR is highly similar
, although there appears to be significant differences between activation kinetics and cytokine production capacity of CAR T cells, including CD28 or 4-1BB signaling domains, in some studies.
Although CD28 CAR T cell-induced CRS tends to predate 4-1BB CAR T cells, cytokine profiles in patient serum differ little in peak levels of cytokines and chemokines, consistent with
their common pathophysiology 。 IL-6, IL-8, IL-1 receptor antagonists (IL-1Ra), CC-chemokine ligand 2 (CCL2), and CCL3 (not major T cell derivatives) are generally elevated, suggesting that CRS may have a common mechanism
beyond the CAR T cells themselves and involving host cells.
A schematic of the pathophysiology of CRS is shown in Figure 2
.

The similarities between CRS and systemic inflammatory response syndromes such as hemophagocytic lymphohistiocytosis and macrophage activation syndrome are discussed
in BOX 1.
Although these and other late-onset inflammatory toxicities have been observed to be independent of CAR T cell specificity, data suggest that CD22CAR T cells (developed to target tumors secondary to CD19 antigen escape) occur more
frequently.

Cell interactions and molecular mediators
.
Both mouse models ultimately showed that CRS was caused by a multicellular network involving CAR T cells and host cells, with macrophages and monocyte cell line cells being the most important
.
One study showed that in the CD19+ lymphoma xenograft model, multiple endogenous cell populations at the tumor site, including dendritic cells, monocytes, and macrophages, produced IL-6 and macrophages far outperformed other cell types
.
In this allogeneic case, CAR T cells produce significantly higher levels of IL-6 in mice than human IL-6
.
IL-6 production is not induced at a distal tumor-free site, supporting the idea that
CRS originates locally but produces systemic pathology.
Another study showed that humanized NGS mice using patient-derived leukemia cell lines (macrophages depleted with clodronate), or given CAR T cell-mediated targeting prior to therapeutic CAR T cell administration, eliminated IL-6 production and significant CRS
.
Leukocyte single-cell RNA sequencing data isolated during CRS confirm that monocyte cell line cells are the origin
of IL-6.
In both models and based on clinical experience, blocking IL-6 maximizes CRS-related toxicity
.
Both studies further revealed that IL-1, a cytokine of monocyte and macrophage origin, is also a potent driver of CRS-associated toxicity
.

Potential triggers for macrophage recruitment or activation during CRS are emerging
.
CAR T cells themselves must be activated to induce myeloid cells to produce cytokines, consistent with clinical observations that patients who do not respond to CAR T cell therapy have mild to no
CRS 。 It is uncertain whether contact-dependent interactions are required between CAR T cells and host bone marrow cells, and CD40-CD40L interactions, while not required to cause CRS, can exacerbate macrophage activation, thereby increasing IL-6 production and aggravating CRS; Which other exposure-dependent mechanisms, if any, may be critical to the development or amplification of CRS remains to be determined
.
T cell surface molecules, such as integrin LFA1 and the co-stimulatory molecule CD28, interact with highly expressed homologous receptors in bone marrow cells (ICAM1 and CD80 or CD86, respectively), and may be particularly valuable in CRS based on the potential role of CD28-CD80/CD86 bidirectional signal transduction and induction of IL-6 in bone marrow cells, and it has been confirmed that blocking these interactions may reduce the severity
of CRS.

Cytokine mediators: IL-6 and IL-1
.
As noted above, the contact-dependent interaction of the cytokines IL-6 and IL-1 plays an important role
in the pathophysiology of CRS.
IL-6 is a pleiogenic cytokine with both pro-inflammatory and anti-inflammatory effects, it is mainly produced by macrophages and other cells of the myeloid line, acts in an autocrine manner, and in combination with other inflammatory signals promotes macrophage maturation and activation
.
IL-6R is primarily expressed
by immune cells (including microglia) and hepatocytes.
IL-6 can transmit signals and promote inflammation through cis-signaling and trans-signaling, and trans-signaling has a wide range of roles
outside the immune system when soluble IL-6–IL-6R forms complexes with widely expressed membrane-bound gp130.
For example, IL-6 is involved in hyperthermia, glucose metabolism, neuroendocrine system, and appetite regulation in addition to controlling acute phase reactions, but the exact role of IL-6 in CRS remains unclear
.
Blocking IL-6 leads to most symptom reversal and full cytokine downregulation in many patients, and preclinical models have also found that IL-6 modulates CRS mortality and promotes macrophage activation by inducing nitric oxide synthase (iNOS) and nitric oxide (NO) production
.

IL-1 is also a pleiopotent cytokine with multiple functions, produced primarily by monocytes and macrophages
.
The IL-1 receptor (IL-1R1) is widely expressed and is responsible for transducing pro-inflammatory signals
.
IL-1 induces tissue production of downstream pro-inflammatory cytokines (e.
g.
, IL-6) and a range of chemokines that can organize mature cells and recruit immune cells, activates pro-inflammatory lipid mediators (e.
g.
, prostaglandin E2, which promotes edema), induces acute-phase proteins and signals the hypothalamus to induce fever, and signals to the pituitary and adrenal glands, with direct and indirect effects
on the circulatory system.

IL-1 was found to be a key medium for CRS in two independent studies (Figure 2): blocking IL-1-induced signal transduction with the IL-1R antagonist Anakinra protected mice from weight loss and fever and prevented CRS-related death in a humanized xenograft model; In a SCID–beige xenograft model, anakinra protects mice from CRS-associated death and reduces macrophage expression of iNOS; CAR T cells that have been engineered to express IL-1Ra also provide protection against the lethality of CRS
.
Importantly, in both mouse models, the antitumor efficacy of CAR T cells remained intact and CRS
was inhibited simultaneously when IL-1 inhibition was initiated prophylactically.
The duration of treatment for anti-IL-1 therapy may be critical
to its efficacy.
Interestingly, in humanized NSG mouse models, blocking IL-6 did not improve macrophage infiltration into the brain compared to blocking IL-1R
.

Based on animal studies, several clinical trials have been initiated to evaluate the role of anakinra in CRS and neurotoxicity prevention (ClinicalTrials.
gov NCT04148430, NCT04205838, NCT03430011, NCT04432506, and NCT04359784).

。 Given the clinical accessibility of IL-1β-blocking agents such as the monoclonal antibody canakinumab, future preclinical studies should shed more light on the specific role of IL-1α and IL-1β in the CRS cascade, as these cytokines show large overlap in pro-inflammatory function, but can also be distinguished
by differences in expression and disease background.

Damage-related molecular patterns and other soluble mediators
.
CAR T cells tend to induce cell death through the inflammatory process of pyroptosis rather than apoptosis, resulting in the release
of damage-related molecular patterns such as ATP and HMGB1.
Molecular patterns associated with damage to tumor cell release lead to in vitro macrophage activation to produce IL-6 and IL-1, whereas inhibition of in vivo pyroptosis in the SCID–beige xenograft model reduces CRS-related mortality (but may also affect cytolytic activity of CAR T cells).

Granulocyte-macrophage colony-stimulating factor (GM-CSF) is produced by a variety of cells, including activated CAR T cells, and blocks the elimination of monocytes to produce IL-6 and other cytokines
in vitro.
However, IL-6 production is extremely powerful, CRS is ultimately fatal in xenograft models, and human T cell-derived GM-CSF does not cross-react with homologous mouse receptors, but GM-CSF may still be a contributing factor to CRS and CRS
is not needed.
In the human PBMC xenograft model, blocking both mouse and human GM-CSF reduces IL-6 production (Figures 2, 3).

Tumor necrosis factor (ΤΝF) is another pleioactive cytokine that mediates inflammation by activating T cells, macrophages, monocytes, and other cells of the immune system, while TNF, unlike IL-1 and IL-6, also has direct tumor-killing activity
.
Recently, it has been proposed in a HER2 humanized mouse mammary tumor virus (MMTV) breast cancer model that TNF is a potential factor
in CD3 and HER2 bispecific antibody-induced CRS.
In this model, TNF inhibitor pretreatment reduces circulating IL-6 (and in some cases, IL-1) without compromising antitumor efficacy
.
In the SCID–beige xenograft model, blocking TNF significantly reduces myelocyte production of IL-6 and eliminates CRS-related mortality, but also reduces the antitumor activity of CAR T cells, depending on the CAR structure (unpublished data) (Figures 2, 3).

Interferon-γ (IFNγ) is produced in large quantities by activated T cells and enhances NO production by promoting macrophage maturation through iNOS induction, and can also promote increased
permeability of other tissues, including the blood-brain barrier, by loosening tight junctions 。 Serum concentrations of IFNγ can increase up to 100-fold in patients with severe CRS, also in macrophage activation syndrome (BOX 1), and despite this observation, the specific causative role of IFNγ in CRS has not been reported
.

The adrenal system is also involved in the occurrence and maintenance of CRS, as catecholamine epinephrine and norepinephrine directly affect CAR T cell activation and subsequent cytokine release
.
Administration of atrial natriuretic peptides (ANPs, regulators of electrolyte and extracellular fluid volume and inhibitors of cytokine secretion) or inhibition of catecholamine synthesis by methtyrosine reduced cytokines produced by CAR T cells in vitro and in vivo, and also reduced mortality in xenograft mouse models (Figures 2, 3).

In a homologous mouse model of B-cell acute lymphoblastic leukemia, low-dose methyltyrosine did not impair antitumor efficacy
.
Interestingly, immune cells are not only activated by catecholamines, but when activated, they naturally produce catecholamines, creating a positive feedback loop
.
However, ANP dysregulation promotes edema, decreased blood pressure, and electrolyte imbalances, all of which are common
in CRS pathology.
Therefore, blocking catecholamine receptors is safe and feasible, especially in the setting of CRS, which is of concern
.

Pathophysiology of ICANS

Although the clinical features of ICANS are easy to identify, its pathophysiology remains poorly
understood.
Recent animal models have shown that endothelial cell activation and disruption of the blood-brain barrier can directly lead to neuronal cell damage in addition to the effects of various pro-inflammatory cytokines, but these models of CAR T cell-mediated neurotoxicity are limited
by the lack of human cytokines and hematopoietic cells and the occurrence of concomitant xenograft-versus-host disease.
Nevertheless, recent studies in mice and non-human primates have yielded important insights that, together with data from clinical trials of ICANS following infusion of CAR T cells, have improved our understanding of the pathophysiology of ICANS (Figure 4).

Vascular permeability, endothelial destruction, and glial damage
.
Patients with ICANS have elevated levels of proteins, CD4+ T cells, CD8+ T cells, and CAR T cells in CSF, indicating loss of blood-brain barrier integrity; Clinical studies have shown that both the number of CAR T cells and the level of cytokines in CSF correlate
with ICANS severity.

Some of the biochemical features of severe CRS and ICANS, such as hypofibrinogenaemia and increased fibrin degradation products, are the same as those of disseminated intravascular coagulation and endothelial cell destruction, which is typically seen in sepsis and critical illness
associated with increased vascular permeability.
In fact, there is evidence of vascular leakage
in patients with severe ICANS.
In a single-center study of 133 patients receiving CD19CAR T cell therapy, changes
in the angiopoietin-TIE2 axis that healthily regulate endothelial cell activation were described.
Platelets and perivascular cells produce ANG1, which stabilizes the endothelium when bound to its endothelial
receptor, TIE2.
Upon activation of endothelial cells by inflammatory cytokines, ANG2 is released from the endothelial Weibel–Palade body and replaces ANG1, further increasing endothelial cell activation and microvascular permeability
.
Consistent with this mechanism, patients with severe ICANS have a statistically significant increase in the ratio of serum ANG2 to ANG1 compared with less neurotoxic patients, along with elevated concentrations of von Willebrand factor (vWF) and CXC-chemokine ligand 8 (CXCL8), both produced
by platelets and perivascular cells.
In addition, patients with elevated serum ANG2 to ANG1 ratios before cleansing and CAR T cell administration are at higher
risk of developing ICANS.
Further data suggest that sequestrated high-molecular-weight vWF multimers can lead to coagulopathy
in patients with severe ICANS.
However, the factors that control baseline and interfere with ANG2 and ANG1 levels remain unclear
.

In addition to BBB disruption and increased vascular permeability, glial damage
has been reported in children and young adults who develop ICANS after CD19CAR T cell therapy.
In a cohort of 43 patients, GFAP and S100b levels were significantly elevated
in CSF in patients with acute neurotoxicity.
GFAP is a well-validated marker of astrocyte injury regardless of cause, while S100b in CSF is a marker of
astrocyte activation.
In addition, this study found that elevated levels of IL-6, IL-10, IFNγ, and granzyme B in CSF were also associated with
neurotoxicity.

Cytokines and their cellular sources
.
In humanized NSG mouse models, infusion of human CD19CAR T cells carrying CD28 or 4-1BB co-stimulatory domains resulted in B-cell aplastic anemia, CRS, and neurotoxicity
.
These mice developed delayed lethal ICANS after initial CRS, which has also been observed in some patients (typically < 1%)
.
However, in this model, blocking signal transduction via IL-6R had no effect on neurotoxicity; In contrast, blocking IL-1R eliminates CRS and neurotoxicity without affecting the efficacy
of CAR T cells.
Monocyte ablation negatively affects
CAR T cell proliferation and population expansion.
This difference may be due to the fact that the IL-1R antagonist anakinra can cross the BBB, while there is no clear evidence that IL-6R-specific monoclonal antibody tocilizumab can penetrate the CNS
.

ICANS models of non-human primates using immunocompetent rhesus macaques have recently reported that these primates developed characteristics consistent with ICANS after infusion of CD20CAR T cells carrying a 4-1BB co-stimulatory domain
.
Although an increase in the number of CAR T cells and non-CAR T cells in both CSF and brain parenchyma was observed during peak neurotoxicity, the number of CAR T cells was higher than that of non-CAR T cells
.
This is associated with high concentrations of IL-6, CXCL8, IL-1Rα, CXCL9, CXCL11, GM-CSF, and vascular endothelial growth factor, which are higher in CSF than in the corresponding serum samples
.
Varying degrees of histological panencephalitis, including multifocal meningitis and perivascular T cell infiltration, are seen 8 days after CAR T-cell infusion, coinciding
with peak levels of CAR T cell proliferation in peripheral blood.
Detailed phenotypic analysis found a significant increase in the expression of VLA4 on the surface of CAR T cells compared with non-CAR T cells, possibly promoting increased
transport of CAR T cells to the CNS.
Taken together, these findings suggest that the mechanisms driving ICANS may include pro-inflammatory cytokines and accumulation of CAR T cells in the CNS, but their relative contributions have so far been unclear
.

Numerous clinical trials have shown that elevated serum levels of multiple cytokines are associated
with the risk of developing ICANS.
Persistently increased cytokines have been found in multiple studies using multiple CAR structures and different target malignancies in patients including IL-2, IL-6, IL-10, and IL-15
.
However, even though mouse models have shown a well-defined role in receptor monocyte-derived immune cells in cytokine secretion and CRS and ICANS pathogenesis, specific cellular sources
of cytokines in ICANS patients have not been identified in clinical trials.
However, a significant increase
in the number of bone marrow cells was observed in CSF in patients with severe ICANS.

Target antigen expression
in CNS.
Clinical trial data suggest that the presence of antigen-positive tumor cells
in the CNS is not required for ICANS to occur.
It is also important to note that ICANS
does not occur in patients with CAR T cells when intrathecal or intratumorous are infused in patients with glioblastoma multiforme.
However, in a recent study, single-cell RNA sequencing analysis demonstrated that CD19 is expressed in human brain parietal cells, including pericytes and vascular smooth muscle cells, so the on-target off-tumour effect may be associated with
CD19CAR T cell neurotoxicity.
This observation may explain why CD19-targeted therapy has a higher incidence of ICANS compared with treatments that target CD20, CD22, and BCMA (also known as TNFRSF17), but alternative or other mechanisms
may also be present.
CD22 expression by human brain microglia was not associated
with a higher incidence or severity of ICANS during CD22CAR T-cell therapy.
Microglia are special phagocytes that can persist in the CNS for decades; They engulf myelin fragments and protein aggregates, thereby preventing neuronal cell damage and maintaining brain homeostasis and function
.
Using CRISPR–Cas9 knockout screening in conjunction with RNA sequencing, the role of CD22 as a negative regulator of microglial phagocytosis has been identified (previously unreported).

CD22 is upregulated in elderly microglia mice, thereby impairing the clearance of myelin fragments, β-amyloid oligomers, and α-synuclein fibrils in vivo, while administration of CD22-blocking antibodies reverses microglial dysfunction and improves phagocytosis and cognitive function
。 The role of microglia phagocytosis in the pathogenesis of ICANS and the consequences of microglial expression of CD22 on CD22CAR T cell therapy remain elucidated, but cytokine-mediated microglial activation has been reported in children with cerebral malaria who have diffuse encephalopathy with BBB destruction and cerebral oedema, with clinical features no difference
from ICANS.

Cerebral edema
.
Cerebral oedema is a rare but potentially fatal neurological complication
following CAR T-cell therapy.
Available evidence suggests that the pathophysiology of cerebral oedema may differ from
the more common encephalopathy manifestations in ICANS.
In a clinical trial evaluating CD19CAR T-cells in adults with B-cell acute lymphoblastic leukaemia, five patients developed fatal cerebral oedema leading to trial termination
。 Root cause analysis assessing patient characteristics, preconditioning treatments, and product attributes showed that patients with cerebral edema developed were < 30 years of age, a higher percentage of CD8+ T cells in CAR T cell products, high serum IL-15 levels and low platelet levels before CAR T cell infusion, and rapid expansion of the CAR T cell population, peaking within week 1, associated with a sharp increase in serum IL-2 and TNF levels
.
Importantly, autopsies in 2 patients showed complete destruction of BBB but no activated T cells
in the CNS.
Although these results are not clear, they suggest that disruption of the blood-brain barrier and subsequent cerebral edema may be due to a surge in inflammatory cytokines rather than infiltration of CAR T cells into the CNS
.
Furthermore, the analysis suggests that cerebral oedema may be due
to a multifactorial confluence that includes patient characteristics and product attributes.

Clinical management of CRS and ICANS

Low-grade CRS can be managed with supportive care with antipyretics, but ensure that no other factors contribute to fever (e.
g.
, infection).

Moderate/severe CRS can be treated with IL-6R-blocking antibody tocilizumab with/without glucocorticoid immunosuppression, and intensive supportive care, including fluid resuscitation and vasopressors for hypotension, and supplemental oxygen
as needed for hypoxia.
Low-grade ICANS are also usually managed with diagnostic tests and supportive care, while severe ICANS is usually treated
with corticosteroids in most centers.
Tocilizumab significantly reduces the incidence of severe CRS, possibly due to the fact that IL-6 peaks early in CRS, and IL-6 is a key mediator
of the downstream inflammatory cascade.

While tocilizumab is very effective for CRS, it is not effective in most ICANS, possibly due to pathophysiology differences between CRS and ICANS and/or poor
permeability of tocilizumab across the BBB.
In fact, prophylactic tocilizumab reduces the incidence of severe CRS but increases the incidence of severe ICANS, possibly due to elevated serum IL-6 levels following tocilizumab administration, which is due to receptor blockade preventing uptake into peripheral tissues
.
However, these observations need to be interpreted with caution because the studies were not random and the sample size was small
.
Although corticosteroids are available through the BBB and are commonly used in ICANS, there is a lack of definitive evidence
of clinical benefit in ICANS severity or duration.
Most studies suggest that the use of tocilizumab does not appear to affect the efficacy of CAR T cells; Data on the efficacy of corticosteroids on CAR T-cells are conflicting, with some studies suggesting no effect, but others showing adverse clinical outcomes with an increased
risk of early progression and mortality.
This contradiction further underscores the need to better understand the pathophysiology of CRS and ICANS, and to develop new strategies
for management.

Influence of
other clinical factors.
The onset, severity, and duration of CRS and ICANS after CAR T-cell therapy may be influenced
by host, tumor, and/or treatment-related factors.
Patients with a higher baseline inflammatory status (defined as C-reactive protein, ferritin, D-dimer, and pro-inflammatory cytokine levels) are at increased risk of CRS and ICANS, who may have a predisposition to
severe inflammatory response after CAR T-cell infusion.
In addition, severe cases of CRS and/or ICANS are associated with a higher tumor burden in a variety of malignancies, including leukaemia, lymphoma, and multiple myeloma, which may be associated with
larger and synchronized activation of the CAR T cell population 。 The intensity of pre-infusion pretreatment therapy for CAR T cells also appears to influence toxicity severity, and stronger regimens may increase the risk of severe CRS and/or ICANS by inducing larger showers, which eliminate cytokine sinks and enable higher levels of homeostatic cytokines such as IL-2 and IL-15, which can be used for CAR T cell proliferation
.
Interestingly, advanced age does not increase the risk of
severe CRS or ICANS.

Impact of
CAR design.
The design of the CAR molecule itself can significantly affect the proliferation and cytokine profile of CAR T cells, thereby affecting the incidence and severity
of CRS and/or ICANS.
As mentioned above, CAR T cell products carrying the CD28 co-stimulatory signal domain appear to proliferate faster after infusion, and their numbers appear to peak
earlier than those engineered with the 4-1BB co-stimulatory signal domain.
In clinical studies, this has been associated with
earlier onset of CRS and ICANS and a higher incidence of severe CRS and ICAN.
However, to date, there has been no direct comparison
of CAR T cell products in the CD28 and 4-1BB co-stimulatory domains.
In addition, differences in patient characteristics (e.
g.
, tumor burden) and differences in toxicity monitoring and grading between studies may also explain some differences
in pharmacokinetics and adverse events in the CD28 and 4-1BB domains.

Literature evidence also suggests that altering the non-signaling domains of CAR molecules, including hinge and transmembrane regions or antigen-binding domains (scFv), can also affect CAR T-cell therapy-related toxicity
。 For example, in preclinical models and phase I clinical trials, changing the length of the CD8α hinge and transmembrane domain reduced cytokine production and reduced proliferation of CD19CAR T cells carrying 4-1BB and CD3ζ signaling domains, while still retaining their cytolytic activity; It is important to achieve a high complete response rate in patients with B-cell lymphoma, with only low-grade CRS and no ICANS
.
CD19CAR T cells carrying low-affinity scFv have higher antitumor efficacy in children with acute lymphoblastic leukemia but do not induce severe CRS
compared with the study of Tisagenlecleucel, a CD19CAR T-cell therapy using high-affinity scFv.

Patients with B-cell lymphoma treated with CD19CAR T cells with CD28 and CD3ζ signal transduction domains (same as approved CD19CAR T cell axicabtagene ciloleucel but different scFv, hinge, and transmembrane domains) had similar antilymphoma activity, but the incidence and severity of ICANS were significantly lower
。 In a small study of CD19CAR NK cells, patients with B-cell malignancies achieved high complete response rates without CRS or ICANS, suggesting that different immune effector cells may have different toxicity profiles
.
Finally, CAR encoding a single immune receptor tyrosine-activated motif in CD3ζ (rather than 3 of them) has been shown in preclinical studies to enhance efficacy and memory T cell differentiation without increasing inflammatory activity
.

In summary, the above reports suggest that reducing tumor burden and baseline inflammatory status, adjusting pretreatment regimens, and optimizing the design of CAR molecular and/or CAR T cell products may reduce the incidence and/or severity
of CRS and ICANS.

New policies to manage CRS and ICANS

Hypersecretion and dysregulation of cytokines are at the pathological core
of CRS.
Nonspecific immunosuppressants (e.
g.
, corticosteroids) may relieve patient symptoms in many cases, and other broad-spectrum cytokine inhibitors such as ruxotinib that block JAK1 and JAK2 or itacitinib that blocks JAK1 (a kinase required for cytokine receptor signaling) are expected to attenuate the effects
of pro-inflammatory cytokines such as IFNγ and IL-6.
In fact, in preclinical models of CAR T-cell-associated toxicity, both ruxolitinib and itacitinib reduce toxicity and cytokine secretion
.
Similarly, ibrutinib, a BTK inhibitor, also inhibits IL-2 tyrosine kinase (ITK; A kinase involved in proximal T cell receptor signaling), resulting in decreased cytokine secretion (including IL-6) in mouse models, but also reduced levels of CAR T cell-derived cytokines such as IFNγ, which may indicate decreased
CAR T cell activation.
The ibrutinib study monitored anti-tumor efficacy only in the short term (day 4 after CAR T cell metastasis), so no difference was
observed.
Interestingly, in co-culture with mantle cell lymphoma cell lines, ibrutinib reduces cytokine secretion in tumor cells, demonstrating the versatility of this treatment
.
Small molecule kinase inhibitors often bind to multiple targets, raising the possibility that ruxolitinib and ibrutinib directly affect CAR T cell activation levels and thus clinical outcomes
.

Another study took advantage of the incomplete specificity of kinase inhibitors to create an "on/off" switch
for CD3ζ-chain-based CAR T cells.
Dasatinib is a BCR-ABL-targeted kinase inhibitor approved for the treatment of various hematological malignancies and is capable of rapidly and reversibly potent inhibition of CAR T cell-mediated cytotoxicity and cytokine production; Short-term administration of dasatinib in preclinical models also reduces CRS-related mortality without compromising in vivo antitumor efficacy after drug removal and restoration of CAR T cell activity (Figure 3).

Another recent preclinical study showed that attenuating LCK signaling by modifying the enzyme SHP1 to recruit into immune synapses further supports the important role
of activated CAR T cells in initiating CRS by reducing the production of effector cytokines by CAR T cells as well as the severity of CRS 。 In addition, the hinge, transmembrane, and proximal intracellular domain structures of BBz CAR (carrying 4-1BB and CD3ζ co-stimulatory domains) were modified to present a phenotype of passivation activation and reduced production of effector cytokines, and toxicity was reduced
in SCID–beige xenograft preclinical models and clinics compared with historical data.

Overall, these studies suggest that broad-spectrum cytokine inhibition (through different mechanisms) reduces CRS pathology, a result that is to be expected
based on our understanding of the pathophysiology of CRS.
However, long-term inhibition of cytokines may be detrimental to the antitumor efficacy of CAR T cells, so targeted interventions aimed at selectively disrupting specific cytokine signaling pathways may be preferred
.
In a small case series (n=8) in patients with severe ICANS or hemophagocytic lymphohistiocytosis after CD19CAR T-cell therapy, anakinra blocking IL-1R appeared to benefit half of the patients
.
The judicious use of newer drugs targeting other cytokines (e.
g.
, IFNγ, emapalumab) may help elucidate their role in the management of severe CRS, and future clinical studies should demonstrate the benefits
of such approaches in prevention and/or treatment.

Conclusion and future direction

The current widespread adoption of CAR T-cell therapy is limited in part by the need for centers with experience managing common toxicities such as CRS and ICANS, and the resulting economic and health burden
.
A deeper understanding of the molecular and cytopathophysiology of CRS and ICANS will facilitate the development of effective targeted therapies to reduce toxicity without compromising antitumor activity
.
In addition, the newly designed CAR structure can also minimize the risk of causing CRS and ICANS, while optimizing tumor antigen recognition and efficient T cell signaling
.

Another current limitation is the need to learn more about the biology and mechanisms of action of CAR T cells, including the biophysical properties of CAR and the effects of its co-stimulatory domains on gene expression profiles, potentially altering T cell subsets, function, memory potential, and depletion
.
To improve the clinical efficacy of CAR T-cell therapy, there is an urgent need to generate T cells
with optimal in vivo health, durability, and effectiveness.

In addition to oncology, there is significant interest in the use of CAR-based techniques to generate antigen-specific regulatory T cells with the potential for targeted delivery of immunosuppression
in the areas of autoimmunity and solid organ transplantation.
The in vivo biology and function of these regulatory CAR T cells will differ from the currently observed properties of effector CAR T cells, and some new toxicity
may also occur.
The safety and widespread use of engineered T-cell therapies in oncology and non-malignant indications remains dependent on effective and reliable prevention
of these complications.

References

Emma C Morris, Sattva S Neelapu, Theodoros Giavridis, Michel Sadelain.
Cytokine release syndrome and associated neurotoxicity in cancer immunotherapy.
Nat Rev Immunol .
2022 Feb; 22(2):85-96.
doi: 10.
1038/s41577-021-00547-6.