Mechanism of engineering and historic basing trials for CAR-T cell therapy

Chimeric antigen receptor T cell (CAR-T) immunotherapy is a new autologous cellular therapy that has been developed as an antitumor treatment. Its indications and the number of eligible patients have dramatically expanded over the past decade. Patient’s T cells from peripheral blood are engineered ex vivo with a recombinant T cell receptor (TCR) or a chimeric antigen receptor (CAR), which mediates antibody-targeted recognition and enhances T cell function upon binding [1]. CARs are synthetic receptors consisting of an antigen-binding domain-like extracellular single-chain variable fragment (scFv), transmembrane (TM), and an intracellular domain with tyrosine-based activation motifs (ITAMs) and co-stimulatory signal. The intracellular parts may be different and define five generations of CARs, which are summarized in Fig. 1 [2, 3].

Fig. 1
figure 1

Five generation of CAR-T (adapted from [2, 3])

Full size image

For B cell malignancies, CARs generally bind to CD19 targets and redirect the patient’s own cells to kill tumour cells in 3 main steps (Fig. 2): (1) the antigen-binding domain of CARs recognizes the CD19 antigen on the B cell; (2) the CD3ζ chain signalling domain induces T cell activation and secretion of cytokines; and (3) the co-stimulatory domains increase T cell activation and enhance the cytolytic function [4].

Fig. 2
figure 2

Antitumour mechanisms of CAR-T and cytokine release results in bystander activation of other immune cells (adapted from [7])

Full size image

Based on phase 2 or 3 trials, anti-CD19 CAR-T cells have demonstrated efficacy in the treatment of paediatric and adult acute lymphoblastic leukaemia (ALL), adult refractory or relapsed high-grade B cell non-Hodgkin lymphoma (NHL) (for diffuse large B cell lymphoma (DLBCL), primary mediastinal B cell lymphoma and mantle cell lymphoma), indolent B cell NHL (for follicular lymphoma). Anti-BCMA (B cell maturation antigen) CAR-T cells have demonstrated efficacy in the treatment of multiple myeloma (MM) [5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22] (summarized in Table 1).

Table 1 Development of clinical trials of CAR-T cell therapy
Full size table

Currently, six commercially products have been approved by the Food and Drug Administration (FDA) for adult patients: Two autologous anti-BCMA CAR-T cell products (idecabtagene vicleucel and ciltacabtagene autoleucel) and four autologous second-generation anti-CD19 CAR-T cell products (tisagenlecleucel, lisocabtagene maraleucel, axicabtagene ciloleucel, and brexucabtagene autoleucel). They differ in the co-stimulatory domain (4-1BB for tisagenlecleucel and lisocabtagene maraleucel and CD28 for axicabtagene ciloleucel and brexucabtagene autoleucel) and by the transduction vector (lentivirus for Tisagenlecleucel and Lisocabtagene maraleucel and retrovirus for Axicabtagene ciloleucel and Brexucabtagene autoleucel). Thus, the expansion speed and duration of action differ between products, ranging from weeks for axicabtagene to months for tisagenlecleucel [23].

The whole process of treatment with anti-CD19 CAR-T cells includes patient’s selection, determining eligibility, leukocyte apheresis, and bridging therapy to stabilize the disease and prevent rapid progression during the 3–8 weeks of the cell manufacturing process, which is the vein-to-vein time between leukapheresis and infusion. This is followed by lymphodepletion conditioning and CAR-T cell infusion, after which complications may occur [24].

This short pragmatic review for intensivists focuses on short-term (admission to day 28) and medium-term (day 29–100) complications, including severe life-threatening toxicities possibly requiring admission to intensive care unit (ICU). Management methods for these complications were developed based on the current literature and recent recommendations derived from a comprehensive review on the topic from the European Society for Blood and Marrow Transplantation (EBMT), Joint Accreditation Committee ISCT-Europe (JACIE), and European Haematology Association (EHA) [25].

Short-term complications

Tumour lysis syndrome (TLS)

TLS has been reported in 5–17% of CAR-T recipients [26] and is characterized by hypocalcaemia, hyperkalaemia, metabolic acidosis, hyperphosphatemia, hyperuricemia, and renal failure. TLS should be prevented and managed with adequate monitoring and standard care, including control of potassium and phosphorus intake during the risk period, hyperhydration, and reducing the level of uric acid (with allopurinol or rasburicase). Despite optimal care, severe acute kidney injury (AKI) remains a frequent complication of TLS [27] and may require dialysis according to the AKI guidelines [28].

Infections and sepsis

CAR-T cell recipients have high risk of sepsis, which is one of the main reasons for ICU admission. A high proportion of patients who receive CAR-T cell therapy develop typical bacterial (20%), viral (5–10%), and fungal (< 5%) infections within the first 28 days after infusion [29]. Most of these infections (80%) occur within the first 10 days, and most patients present with lower-respiratory tract infections.

Risk factors for infection after CAR-T infusion include neutropenia, previous antitumor treatment regimens, the CAR-T cell dose, high grade of cytokine release syndrome (CRS) or immune effector cells associated neurotoxicity syndromes (ICANS), and their treatments. Because long-lived plasma cells do not express CD19, so humoral immunity to viruses is preserved, and the occurrence of severe viral infections remains rare with anti-CD19 CAR-T [30]. To date, few studies have specifically addressed the issue of viral infections or reactivations in patients receiving anti-BCMA CAR therapy. However, Wang et al. recently reported that viral infections or reactivations due to double-stranded DNA viruses like herpes virus, adenovirus, and BK or JC viruses were common adverse events in patients receiving anti-BCMA [31].

There are no standardized approaches to antimicrobial prophylaxis regimens for CAR-T cell recipients. Fever after lymphodepletion and CAR-T infusion requires, however, prompt empiric antimicrobial therapy, because infections and sepsis are an important determinant of increased morbidity and mortality [32].

Cytokine release syndrome (CRS)

CRS is the most common acute toxicity induced by CAR-T cell therapy. It is characterized by systemic inflammatory reaction (a “cytokine storm”) with flu-like symptoms, hypoxemia, and haemodynamic instability. It is staged into 4 grades according to consensus criteria of the American Society for Transplantation and Cellular Therapy (ASTCT) [33]. Pathophysiologically, CRS leads to the release of effector cytokines which activate the monocyte/macrophage system and induce the production of pro-inflammatory chemokines. In preclinical models, the main cytokine with the highest concentration is IL-6, which explains the first-line use of the anti-IL-6 receptor tocilizumab for CRS.

In a recent review, the incidence of CRS grade > 2 was reported in 29% of treatments of ALL and 20% of treatments of refractory or relapsed high-grade B cell NHL [42, 49]. Their management is based on etiological treatment and systemic corticosteroids with or without etoposide combination therapy [43].

Cardiovascular toxicity

Cardiovascular complications are reported in 10–20% of CAR-T cell recipients. Risk factors include CRS grade > 1, disease burden, pre-existing cardiac dysfunction, and exposure to cardiotoxin therapy, such as anthracyclines or tyrosine kinase inhibitors [50]. Currently, there are no formal guidelines for risk stratification. Nevertheless, in a recent review, Gutierrez et al. reported a group of patients with high risk for cardiac comorbidities before CAR-T cell infusion including prior or current cardiomyopathy, heart failure with reduced left ventricular ejection fraction (< 50%), prior history of myocardial infarction or coronary revascularization, significant valve disease, and age > 65 years [51].

The mechanisms involved in cardiovascular dysfunction are thought to be primarily mediated by the systemic inflammation of CRS, particularly IL-6. In a recent trial, CAR-T-related severe cardiovascular events were independently associated with increased non-relapse mortality and overall mortality risk [52]. ICU management is not specific, but cardiac MRI has emerged as an interesting tool for the diagnosis of CAR-T cell-related cardiotoxicity and differential diagnoses [53, 54]. Because of the close interaction between CRS and CAR-T cell-related cardiotoxicity, cardiovascular complications must be managed with intravenous tocilizumab and are associated with rapid improvement [52].

Kidney toxicity

AKI is frequent after CAR-T cell therapy, with an estimated incidence of 18.6% [55, 56]. Several mechanisms can explain AKI after CAR-T cell infusion, such as vasodilatory shock after CRS, sepsis, immunoallergic tubulointerstitial nephritis, and TLS. Like any patients with haematological malignancies, AKI and dialysis are strongly associated with increased mortality [57]. Nevertheless, there is no specific management for AKI after CAR-T cell therapy, and dialysis modalities may rely on AKI guidelines [28].

Medium-term complications

Delayed TLS, CRS, and ICANS

All the major short-term syndromes described thus far may occur later and should be managed in the same way.

B cell aplasia and hypogammaglobulinemia

Most of the antigens targeted by CAR-T cell therapy are not exclusively specific to tumours but are also expressed by non-malignant tissues (off-tumour and on-target toxicity). Anti-CD19 or anti-BCMA CAR-T cells target B cell CD19 or BCMA antigens, respectively, so patients can develop B cell aplasia and profound hypogammaglobulinemia. These adverse effects were reported in 25% of cases at 12 months in the ZUMA-1 trial and associated with sino-pulmonary infections [10]. Intravenous immunoglobulins (0.4 g/kg/month) or subcutaneous immunoglobulins (0.1 g/kg/week) are the standard treatment for hypogammaglobulinemia below 4 g/L associated with recurrent infections. Discontinuation of immunoglobulin administration should be guided by the recovery of functional B cells. Notably, 65% of patients receiving CAR-T cell recovered a normal level of absolute B cell numbers with a median time of 12 months (range 2–59 months) [58].

Delayed and prolonged cytopenias

Delayed haematological toxicity may affect up to 65% and increases morbidity and mortality after CAR-T treatment. Several mechanisms can explain prolonged and late cytopenia, such as hyperinflammatory syndrome like IEC-HS, immune-mediated hematopoietic stem cell suppression, mature blood cell destruction, transplant-associated thrombotic microangiopathy, primary disease relapse, and secondary marrow neoplasm. Clinical trials have reported a high incidence of persistent grade > 2 neutropenia (30–40%), thrombocytopenia (20–30%), and anaemia (10–15%) after day 28. In these cases, bone marrow biopsy may be useful to exclude recurrent disease, secondary or non-specific HLH or secondary myelodysplasia [59]. The CAR-HEMATOTOX model is an easy-to-use risk stratification tool of delayed haematological toxicity that was evaluated in 258 patients with refractory or relapsed DLBCL receiving axicabtagene ciloleucel or tisagenlecleucel [60]. This score includes markers associated with the patient’s hematopoietic reserve and systemic inflammatory status prior to lymphodepletion conditioning and injection of CAR-T cells without being predictive of the occurrence of CRS/ICANS/IEC-HS (summarized in Fig. 4). It is associated with a risk of profound and prolonged cytopenias, infectious complications, prolonged hospitalization, and worse clinical outcomes (negative prognostic impact on overall response rate, progression-free survival and overall survival). A score between 2 and 7 is considered high and may indicate antimicrobial prophylaxis in cases of risk factors for sepsis, although there are no strong recommendations.

Fig. 4
figure 4

CAR-HEMATOTOX: to be determined before lymphodepletion to discriminate between a low and a high risk for haematotoxicity, from [60]

Full size image

Platelet and packed red blood cell transfusion support may be necessary, and granulocyte colony-stimulating factor (G-CSF) can be used for severe neutropenia (< 0.5 G/L). In addition, erythropoietin, thrombopoietin agonists, and IEC-HS directed therapy may have a role in these severe situations. Finally, if stem cells from a prior autologous or allogeneic bone marrow transplantation have been persevered and are available for use, a stem cell boost can be used as a last resort in cases of refractory cytopenias [61].

Infections and antimicrobial prophylaxis

In a recent retrospective analysis from the DESCAR-T registry, Lemoine et al. reported the occurrence of late non-relapse mortality after CAR-T cell therapy for DLBCL. In a median follow-up of 12.4 months, most of them were due to infections (52%) [62]. While post-opportunistic infections are bacterial in the first 30 days, viral infections predominate beyond day 30, which mainly occur in the upper- and lower-respiratory tracts. Late reactivation of herpes virus has been reported. Consequently, antimicrobial prophylaxis is warranted until immune reconstitution (summarized in Fig. 3D) [25]. Early and late post-CAR-T fungal infections appear to be rare. In a recent cohort study including 84 patients admitted to the ICU, only 3 (3.6%) developed fungal infections [63].

Characteristics and outcome of patients admitted in the intensive care unit

A few years ago, through the CAR-ICU initiative, a task force of experts in CAR-T cell therapy has launched a practice survey in 11 US hospitals concerning practices for the management of side effects in CAR-T recipients [64]. They recorded CAR products, toxicities, targeted treatment, management practices and interventions in the ICU. The authors highlighted differences between centres in severity criteria on ICU admission for CRS, but not for ICANS. The management of complications in CAR-T patients was relatively consistent between centres, including the use of vasopressors, monitoring of neurotoxicity by electroencephalogram, prophylactic use of antiepileptic drugs and use of tocilizumab. Conversely, other therapies differed between centres, which included fluid resuscitation, mechanical ventilation requirement, and use of corticosteroids. The authors concluded that future studies were needed to homogenize practices and improve the prognosis of patients.

Recent studies have described the epidemiology, treatments, and outcome of multicentre cohorts of adult patients admitted to the ICU for short-term complications induced by CAR-T cell therapy [65,66,67]. Table 2 provides the main characteristics of CAR-T recipients in these observational cohort studies in ICU settings. In summary, the average age of patients admitted to the ICU after CAR-T varied from 57 to 60 years, with a majority being men (from 59 to 66%) and a maximum median SOFA score ranging from 4 to 5. The main indication for CAR-T cell therapy was DLBCL, followed by MM and ALL, which could be refractory or relapsed after 3 to 4 lines of standard chemotherapy. In the CAR-ICU [65] and CARTTAS [66] studies, the number of CAR-T recipients requiring transfer to the ICU for severe toxicity after CAR-T cell treatment ranged from 27 to 35%. In all studies, CRS occurred in around 70% of patients, with the proportion of severe CRS (grade > 2) ranging from 18 to 35% and occurring between 2 and 5 days after CAR-T cell infusion. ICANS occurred in 37 to 75% of patients, with the proportion of severe ICANS (grade > 2) ranging from 15 to 65% and occurring 1–6 days after CAR-T cell infusion. Furthermore, 22–30% of patients had a documented infection, but in the cohort examined by Valade et al., 98% of patients received broad-spectrum antibiotics in the context of neutropenia [67]. Of note, the rate of TLS was not reported, nor were those of cardiomyopathy and secondary HLH, except in the CAR-ICU study where the proportions observed were below 4%. Regarding artificial organ support therapies, almost a third of patients received vasopressors, and around 10% required mechanical ventilation. Less than 5% of CAR-T recipients required renal replacement therapy. Tocilizumab and corticosteroid drugs were used in 60–75% of cases as a first line of treatment in CAR-T-induced CRS, as recommended by the EBMT, JACIE, and EHA [25]. Finally, ICU and hospital mortality varied from 1.5 to 9% and from 12 to 17.5%, respectively.

Table 2 Characteristics of CAR-T recipients in ICU studies
Full size table

Thus, these three retrospective observational studies of patients admitted to the ICU for early complications secondary to CAR-T cell administration show that the population is predominantly male and middle-aged population with an intermediate severity score (SOFA) for acute illness on admission to the ICU, with mainly haemodynamic failure and a relatively low mortality rate.

In the CAR-ICU trial, higher cumulative corticosteroid doses were associated with decreased survival rate, while CRS and ICANS toxicity grades or organ support did not impact the overall survival [64]. In the CARTTAS trial, frailty, bacterial infection, and lifesaving therapy within 24 h of ICU admission were identified as independent risk factors of 90-day mortality [66]. Similarly, Valade et al. identified reason for ICU admission (disease progression vs sepsis or CRS), performance status, and SOFA score as determinants of mortality [67]. Finally, mortality appears to be associated with the severity of the acute illness, particularly in patients whose performance status is impaired or whose malignant haematological disease is progressing. Altogether, multidisciplinary management of severe patients requires early recognition of life-threatening toxicity symptoms related to CAR-T cell therapies, rapid and maximal treatment of organ failures and infections, as well as perfect knowledge of the treatments specific to the management of CRS or ICANS, according to the evidence-based medicine and the international recommendations [25, 68]. Carefully selecting eligible patients and develo** individualized patient management plans are required to improve the prognosis of these serious patients in the era of this new cell therapy with increasingly broad indications.

Challenges and future issues

New perspectives and improvement of CAR-T cell therapy

New CAR-T cell-based therapies continue to be developed and could prove beneficial for other B cell neoplasias. For example, anti-CD30 CAR-T cell recently demonstrated efficiency in refractory or relapsed Hodgkin Lymphoma, without neurologic toxicity [69].

Studies have suggested that upon target engagement, CAR-T cell therapy rapidly increases activation markers, including programmed cell death-1 (PD-1). The expression of PDL-1 on tumour cells associated with PD-1 activation on CAR-T led to the hypothesis that blocking this signalling cascade could increase the activation, proliferation, and cytolytic activity of CAR-T cell therapy [70, 71]. Thus, the combination of new immunotherapies may be able to improve treatment efficacy. For example, the phase 1/2 ZUMA-6 trial was designed to assess the value of treating refractory or relapsed high-grade DLBCL with a combination of CAR-T cell therapy and monoclonal antibody targeting PDL-1 [72].

The use of allogeneic CAR-T cells from living donors is another approach that could change the therapeutic landscape of CAR-T cell therapy. The potential expected benefits are the possible standardization of CAR-T cell products, the possibilities of multiple cell modifications and using an industrialized process to reduce cost, and the immediate availability of these cryopreserved products for patient treatment. In this respect, the phase 1 ALPHA trial was designed to evaluate the benefit of allogeneic CAR-T cell therapy (ALLO-501 and ALLO-647™) in the treatment of refractory or relapsed high-grade DLBCL or follicular lymphoma [73]. Similarly, Mailankody et al. reported the feasibility and safety of allogeneic anti-BCMA CAR-T cell therapy for refractory or relapsed MM [74].

Resistance or relapse after CAR-T cell therapy can be explained by a mechanism of target repression (i.e. loss of CD19 expression). Thus, the creation of autologous CAR-T cells targeting two antigenic profiles, CD19 and CD22, represents an innovative approach to counteract the acquisition of tumour cell resistance to CAR-T through loss of the mono-antigenic target. The efficacy of bispecific CAR-T cell has been recently tested in patients with refractory or relapsed ALL [75], and with refractory or relapsed high-grade DLBCL [76].

Finally, studies using new CAR-T cell strategies are underway for several haematological malignancies to challenge the monopoly of commercial autologous CAR-T. Their main objectives are to improve response rates, avoid the acquisition of resistance, minimize adverse effects, and reduce manufacturing time (a recent trial have used YTB323 or rapcabtagene autoleucel, an autologous CD19-directed CAR-T cell generated by an innovative platform that produces CAR-T in 2 days [77]).

Future strategies to limit toxicities and improve prognosis

Because high rates of complications have been reported in numerous trials, ranging from 40 to 90% for all grades CRS and from 20 to 65% for ICANS [25], several phase 1/2 studies are warranted to assess new prophylactic or curative treatment strategies, particularly by using granulocyte–macrophage colony-stimulating factor (GM-CSF) or anti-IL-1-R. Current research is also focusing on sparing corticosteroid therapy, which may ultimately be responsible for reduced survival by modifying the engineering of CAR-T cell therapy [78], as well as better haemopathy control prior to treatment [79]. Finally, novel cell products like CAR-natural killer cells or CAR-macrophages may have several benefits over CAR-T cells, without surge of inflammatory cytokines, while lowering the risk of CRS and ICANS and reducing risk of “on-target/off-tumour” toxicity [80].

Although a better understanding of pathophysiology has improved the quality of patient care in the ICU setting and has led to increased in-hospital and overall survival [63, 81], CAR-T cell management is currently based on few recommendations with high levels of evidence. All this could evolve over the coming years as the indications for this immunotherapy are extended to autoimmune diseases and solid cancers [82, 83]. Whether related to the causative disease or to the complications of CAR-T cells, questions remain regarding the intensity of ICU management in case of CAR-T-related severe events.

Conclusion

CAR-T cell therapies are develo** rapidly with an increasingly wide range of indications for haematological malignancies, as well as potential for solid cancers and autoimmune diseases. They have demonstrated satisfying response rates and improved survival rates in patients with refractory or relapsed high-grade B cell NHL. The main CAR-T-specific toxicities are CRS, ICANS, and IEC-HS, while the main non-specific complications are infections. The most severe cases may require admission to the ICU for early management. However, the ICU admission rate, the need for organ support, and mortality tend to decline over the years.