Complications after CD19+ CAR T-Cell Therapy

Simple Summary CD19+ Chimeric antigen receptor (CAR) T-cells are used against CD19+ hematologic malignancies, such as high-grade B-cell lymphoma and acute lymphoblastic leukemia. Since this is a relatively new treatment approach, not all potential side effects are well described, and the underlying pathobiology is often not well defined. Here, we summarize current data on the incidence and the current management of CD19+ CAR T-cell complications. We discuss frequently occurring toxicities and we highlight evidence for the occurrence of rarer side effects affecting different organ systems. In addition, we highlight new findings that shed light on the pathophysiology of CAR T-cell-related complications. Abstract Clinical trials demonstrated that CD19+ chimeric antigen receptor (CAR) T-cells can be highly effective against a number of malignancies. However, the complete risk profile of CAR T-cells could not be defined in the initial trials. Currently, there is emerging evidence derived from post approval studies in CD19+ CAR T-cells demonstrating both short-term and medium-term effects, which were unknown at the time of regulatory approval. Here, we review the incidence and the current management of CD19+ CAR T-cell complications. We highlight frequently occurring events, such as cytokine release syndrome, immune effector cell-associated neurotoxicity syndrome, cardiotoxicity, pulmonary toxicity, metabolic complications, secondary macrophage-activation syndrome, and prolonged cytopenia. Furthermore, we present evidence supporting the hypothesis that CAR T-cell-mediated toxicities can involve any other organ system and we discuss the potential risk of long-term complications. Finally, we discuss recent pre-clinical and clinical data shedding new light on the pathophysiology of CAR T-cell-related complications.


Background
CAR-T cells that target CD19 have become standard treatments of relapsed or refractory hematological malignancies, such as B cell acute lymphoblastic leukemia (ALL) and relapsed or refractory large B cell lymphoma (LBCL). In addition, numerous trials in further tumor entities are under way, making it likely that in the near future, there will be a broader clinical use of CAR-T cells with different antigen specificities.
The safety of CAR-T cells is of major concern, since this new class of antitumor therapy has previously unknown side effects. The complete risk profile of CD19+ CAR T-cells could not be defined in the relatively small initial trials leading to approval of CAR T-cell products. This led to the obligation to document toxicities in any patient receiving commercial CAR T-cell products.

Cytokine Release Syndrome
In response to contact with the target antigen, signaling through the CAR induces CAR T-cell proliferation and cytokine production. This leads to a cascade of immune stimulation involving the activation of macrophages. In a xenogeneic model of CRS using immunocompromised mice injected with human CD19+ lymphoma cells as well as CD19+ CAR-T cells, a critical role of macrophage-derived cytokines, such as IL-6, IL-1, and nitric oxide, has been demonstrated [2]. However, multiple inflammatory proteins are elevated in peripheral blood during CRS, including C-reactive protein, ferritin, interferon-γ, interleukins (IL-1, IL-2Rα, IL-6, IL-8, IL-10, IL-15), monocyte inflammatory proteins, monocyte chemoattractant proteins, and tumor necrosis factor receptors.
An important step of the inflammatory cascade seems to be the activation of endothelial cells in the microvasculature of different organs, resulting in capillary leakage. In patients with severe CRS, angiopoietin-2 and serum von Willebrand factor were significantly increased. Moreover, both factors were predictive of severe CRS when assessed prior to CAR-T cell therapy [3]. Other risk factors for the development of CRS include high CAR T-cell and tumor cell numbers as well as low platelet counts, comorbidities, and early onset of CRS [3][4][5][6]. Moreover, the costimulatory capabilities as well as the antigen binding characteristics of the CAR construct are likely to determine the CRS risk [7]. The reported incidences of CRS are quite variable, with ranges between 30% and 100%, not only depending on the presence of risk factors in the patient cohorts but also on different CRS grading systems used in the literature [8][9][10]. The incidences of CRS reported in the literature are summarized in Table 1. In accordance, the incidence of severe CRS (grades 3-4) has been reported to be in the range from 10-30% [8,9]. Of note, more recent publications report lower incidences of severe CRS and lower rates of intensive care unit treatments, especially in more recent CAR T cell products, which are therefore applied in an outpatient setting [11]. Interestingly, in a recently reported phase 1/2 study, a bicistronic CAR targeting CD19 and CD22 followed by an anti-PD1 showed no induction of ≥ • III CRS [12].
In established CAR T cell products, this fact probably derives from improved management of CRS, namely from an earlier start of effective (steroid) treatment, which was also shown in cohort 4 of the ZUMA-1 study [13].
Recently, CD19-CAR-T cell products have been being tested in indolent B-NHL, such as marginal zone lymphoma and follicular lymphoma, with remarkably high anti-lymphoma activity. However, the frequency and severity of CRS was similar to that in LBCL [14].
The typical onset of CRS is between day + 3 and +6 after CAR T-cell infusion, with clinical manifestations ranging from mild pyrexia through severe cases of multi-organ dysfunction. Grading of CRS is recommended using the ASTCT grading system, which is also supported by the EBMT (Table 2) [9,15]. Pharmacologic management of CRS is adapted to the respective CRS grades [9,15].
Grade 1: Symptomatic treatment (e.g., anti-pyretics and fluids), considering broad-spectrum antibiotics. In case of very high fiver or persistence for more than three days, consider Tocilizumab.
Grade 2: In addition, Tocilizumab 8 mg/kg (max. 800 mg), which can be repeated every 8 h for a maximum of four administrations.
Grade 3: In addition, Dexamethasone 10-20 mg, which can be repeated every 6 h. Grade 4: Methylprednisolone 1000 mg/day in exchange for Dexamethasone. In refractory cases, other immunosuppressive agents, such as Siltuximab (anti IL-6) and Anakinra (anti-IL-1 receptor) [16], as well as cytokine absorption methods, have been used successfully [17]. Grades 3 to 4 CRS are managed in intensive care units and standard approaches for cardiovascular and pulmonary support are used.
Of note, a number of CART complications, which are depicted below, are associated or even induced by higher grades of CRS. Table 3 summarizes prevention and therapy strategies for different CD19+ CAR T-cell related complications. Table 1. Incidence of non-relapse mortality (NRM), cytokine-release syndrome (CRS), and Immune Effector Cell-Associated Neurotoxicity Syndrome (ICANS) in patients included in key clinical trials with commercially approved CAR T-cells (Tisa-cel and Ax-cel).  Anti-bacterial prophylaxis during neutropenic phase PJP and herpes prophylaxis till approximately 6 months post CAR T Antifungal prophylaxis not standard

Immune Effector Cell-Associated Neurotoxicity Syndrome (ICANS)
The pathophysiology of ICANS is not well understood mainly because reliable animal models are lacking and because patient material for research is often not available. However, in peripheral blood, a multitude of inflammatory mediators are elevated, similar to the situation in CRS [21,22]. In ICANS, the vascular pathobiology seems to play a central role, as endothelial activation and multifocal vascular disorder have been found in the brain of patients with fatal neurotoxicity. In addition, biomarkers for endothelial activation were predictive of neurotoxicity even before CAR T-cell administration [23]. Another potentially relevant recent finding is that mural cells, which surround the cerebral endothelium and are critical for blood-brain barrier integrity, express CD19. This may help in explaining the increased frequency of ICANS in patients receiving CD19-targeting immunotherapies [24]. It remains to be determined if future CAR T-cell therapies targeting other antigens will have a lower incidence of ICANS. Currently, the most important risk factor for the development of ICANS is the presence of severe CRS. Other risk factors include a high tumor burden and preexisting neurologic disorders [8,21,22].
The reported incidences are variable in a range between 10% and 50% (summarized in Table 1). Recent data showed a similar frequency of ICAN for other B-cell malignancies, such as indolent B-NHL or mantle-cell lymphoma [14,25].
The time to onset of ICANS-related symptoms is more variable as compared to the onset of CRS and was reported to be in a range of 1-34 days after CAR T-cell infusion [6,18,26,27]. Clinical presentation often starts with a deterioration in handwriting and later on, non-specific symptoms, such as disturbances of attention, agitation, confusion, tremors, headaches, or seizures, may occur. The mental status can be disturbed, ranging from minor changes to severe disturbances (e.g., coma). Regularly, ICANS is reversible; however, infrequently, fatal cases due to hemorrhage or cerebral edema may occur [28]. Grading of ICANS is performed according to the recommendations of ACTCT and EBMT as shown in Table 4 [9,15]. The backbone of pharmacologic management of ICANS is steroids according to the respective ICANS grades [9,15]. Grade 1 ICANS is typically not treated with steroids. Dexamethasone (e.g., 10 mg every 6 h) is generally used in ICANS grade 2-3 whereas high-dose methylprednisolone (e.g., 1000 mg) is often used in grade four ICANS. Steroids are tapered over days to several weeks but there is a risk of recurrence of ICANS [29]. In case of refractory ICANS despite high-dose steroids, there is no standard clinical approach.

Complications after CAR T-Cell Therapy Other than CRS or ICANS
Initial publications derived from the pivotal trials reported a limited number of patients (<100 in most trials) with a limited observation period. However, in the last two to three years, many more patients were treated with CAR T-cells, both with approved drugs and in clinical trials. In the US-based CIBMTR database already >2000 patients and in the EBMT database nearly 1000 patients with CAR T-cell therapy were included (current figures given at cibmtr.org and ebmt.org). This explains why additional complications occurring at later time points or occurring less frequently were reported, including neurologic events, cardiotoxicity, pulmonary toxicity, metabolic complications, secondary macrophage-activation syndrome, prolonged cytopenia, certain infections, immune-related events, and subsequent malignancies. Table 5 summarizes the published literature on the incidence of CAR-T-related complications other than CRS and ICANS.

Late Neurologic and Psychiatric Events
In a recent published report, about 10% of patients surviving CAR T-cell therapy longer than three months had neurological events other than ICANS, including ischemic attacks, peripheral neuropathy, and Alzheimer's dementia [30]. The patho-mechanism behind this association is not known. Furthermore, psychiatric events have been detected in 9% of patients undergoing CAR T-cells in that study. However, 50% of those patients had a pre-existing psychiatric disorder [30]. It is not clear whether such side effects are directly or indirectly associated with CAR T-cells, since pathophysiologic mechanisms for these side effects are unclear and no sufficient control patients were included in those analyses.

Cardiovascular Toxicities
Cardiovascular complications have been initially reported in children treated with CAR T-cells for ALL. In the ELIANA trial, grade 3 toxicities of cardiovascular origin were hypotension, fluid overload, and pulmonary edema in more than 5% of patients [6]. Additionally, cardiomyopathy with left ventricular systolic dysfunction was detected in additional retrospective analyses. However, such complications were reversible in most children weeks to months after CAR T-cells [33][34][35].
In the adult patient population, at least two retrospective analyses were published for approved CAR T-cell products. In one study, major cardiovascular events occurred in 17% of patients till one month after CAR T-cell infusion [36]. In another retrospective monocentric study of 60 consecutive adult patients with LBCL, who were treated either with axicabtagene ciloleucel or tisagenlecleucel, 48 cardiovascular adverse events were observed in 32 patients within one year after infusion [32]. Similar to the cardiovascular toxicities seen in the pediatric population, hypotension and fluid retention were most common. Atrial fibrillation and hypertension were additional cardiovascular side effects in adults. Of note, most cardiovascular events were detected in patients also developing CRS [32,36]. The prevailing patho-mechanism seems to be the exacerbation of pre-existing cardiovascular damage caused by CRS-related stress.

Pulmonary Toxicity
Pulmonary toxicities are complications of special interest in the field of immunotherapies, especially for checkpoint inhibitor therapies. In CAR T-cell therapy recipients, pulmonary toxicities were manageable in most of the cases and no unsuspected lung toxicity occurred to date. However, pulmonary complications are also more frequent in patients with higher grade CRS [32]. The most frequent pulmonary symptom was hypoxia, but also pleural effusion, pulmonary embolism, allergic rhinitis, and pneumomediastinum were described [32]. To date, there is no comprehensive analysis for lung toxicity in recipients of CAR T-cell therapy, especially long-term follow-up with consecutive lung function tests, including transfer capacity of the lung, which are currently missing.

Metabolic Complications
In a recent study with 60 patients by Wudikarn and colleagues, metabolic toxicities after CART were detected as a frequent complication [32]. Typical electrolyte abnormalities were hypokalemia, hypophosphatemia, and hypocalcemia. Hyperglycemia and hypoglycemia were also frequent [19,20,32]. However, such complications are manageable, and persistence is not expected.

Secondary Macrophage-Activation Syndrome (sHLH/MAS)
A recent poll by the EBMT reported an estimated incidence of sHLH/MAS between 3% and 4% [37]. However, it is difficult to dissect secondary macrophage-activation syndrome after CART from severe CRS because some of the diagnostic criteria are overlapping, leading to low practicability of the conventional diagnostic criteria [37]. A reasonable solution that has been proposed recently is that the diagnosis of sHLH/MAS should be made in case of very high ferritin levels (e.g., >10,000 ng/mL) in addition to the typical symptoms [38]. The management of sHLH/MAS after CAR T-cell therapy is currently a matter of debate, but Tocilizumab and steroids are considered the therapeutic backbone. Whether additional agents have positive effects is not yet clear. One opportunity is to administer etoposide in reference to HLH/MAS management outside the setting of CAR T-cell therapy [39]. Another option is to use Emapalumab, an anti-interferon-gamma antibody, which is highly upregulated during MAS/HLH [40].

B-Cell Aplasia
Single infusion with CD19-directed CAR T-cell products induces B-cell aplasia associated with hypogammaglobulinemia, which is shown in numerous studies [6,41]. In most patients, immunoglobulin replacement was used to prevent infectious complications. However, currently, it is not clear whether replacement therapy adds to outcome/improved quality of life in CD19 CAR T-cell patients [42]. Persistence of CD19-directed CAR T-cells will most probably maintain B-cell aplasia. However, long-term studies will clarify whether CD19 CAR T-cell products will be active in the recipient's body for years. Whether B-cell aplasia or hypogammaglobulinemia will be of concern in CART of other B-lineage target epitopes, such as BCMA, will be of interest.

Prolonged Cytopenia
Early, late, and prolonged hematotoxicity after CAR T-cells occur in different CAR constructs and are the most common adverse events after lymphodepletion and CAR T-cell transfusion, with neutropenia being the most frequent. Early cytopenia is frequent in >80% of patients, whereas late cytopenia occurs in 30-40%. In a recent publication on late effects after CAR T-cell therapy, a persisting (after day + 90) cytopenia rate of 16% was described [30]. Overall, between 8% and 18% of CAR T-cell recipients were reported to have prolonged cytopenia. Lymphopenia is seen in 35% of patients at 1 year after anti-CD19 CAR T-cell therapy. One study described that late hematologic toxicity is most frequent in patients with CRS and in patients with a history of stem cell transplantation [43]. In patients with relapsed or refractory large B-cell lymphoma treated with axicabtagene ciloleucel (ZUMA-1 and ZUMA-9 trials), 48% of patients had grade 3-4 cytopenia at day 30 (anemia 16%, neutropenia 29%, and thrombocytopenia 42%). After two years, grade 3-4 cytopenia persisted in one out of nine evaluable patients (11%) [31]. It is largely unknown which factors contribute to cytopenia, e.g., insufficient or late hematopoietic recovery after CART. In one study, 83 patients receiving CD19 CART were evaluated for hematopoietic recovery. In this cohort, baseline cytopenia, higher grades of CRS or ICANS, and higher peak of C-reactive protein or ferritin were associated with a lower likelihood of complete count recovery [44]. Disease entities play a minor role, since cytopenia is observed in different entities, such as myeloma, ALL, and lymphoma. The use of G-CSF might be beneficial in patients without signs of cytokine release syndrome and prolonged cytopenia.

Infections and Vaccinations
It has been reported that in patients surviving day+90 after CAR T-cell therapy, late infections occurred in approximately 60% of patients. Of those infections, 60% were bacterial infections, 31% were viral infections (mostly respiratory viruses), and 9% were fungal infections [30]. In another study, 60 patients with LBCL were retrospectively analyzed for incidence, risk factors, and management of infections [45]. In total, 101 infections were observed in this patient cohort, among them, 23 severe, one life-threatening, and one lethal. The cumulative incidence of infections at one year was 63.3%. Most infections were bacterial (57.2%), severe bacterial (29.6%), viral (44.7%), and fungal (4%). The use of corticosteroids for management of CRS or ICANS was associated with an increased risk of infections. It is recommended to use antimicrobial, antiviral, and antifungal prophylaxis in patients after CART, not only in cytopenic patients. Most centers use Cotrimoxazol Trimethoprim and Aciclovir until approximately 6 months after CD19+ CAR T-cell infusion. However, antifungal therapy, e.g., with Posaconazole, could also be considered.
Data on infection in the context of COVID-19 and CART are sparse. However, it was recently reported that in a cohort of 77 patients with SARS/COV-2 infection receiving cellular therapy (autologous, allogeneic, and CART), a favorable outcome was reported. Overall survival at one month after cellular therapy was 78% [46]. This data suggest that COVID-19 infections in CAR T cell recipients might be manageable. Nevertheless, robust data in a larger cohort is pending.
It is a matter of debate if, and when, patients after CD19 CAR T-cell therapy should be vaccinated as these patients may have low vaccine responses. On the other hand, these patients often have considerable risk for infections and the vaccination efficacy does not exclusively depend on CD19+ B-cell responses. In fact, patients after CD19+ CAR T-cell therapy may have persistence of plasma cells and may be able to have adequate vaccine responses [47]. Therefore, the administration of re-vaccination should be considered in all recipients of CAR T-cell therapy. We suggest performing seasonal influenza A vaccinations, also for close relatives of the patient. Beginning at 6 months after CAR T-cell infusion, a comprehensive re-vaccination boost can be started; however, live attenuated vaccines should not be administered within the first year [48]. Recommended vaccinations are similar to the re-vaccination schedule after allogeneic stem cell transplantation [49]. Of note, the assessment of anti-infectious antibody levels (titers) in patients after substitution of immunoglobulins is misleading and not recommended.

Immune-Related Events and GVHD
In a study on later effects after CAR T-cell therapy, an incidence of 8% of immune-related events was reported, including alveolitis, pneumonitis, persistent skin rash, collagenous colitis, and persistent flu-like syndrome [30]. Generally, it is not expected that autologous or allogeneic donor-derived CAR T cells induce or trigger GVHD or GVHD-like symptoms [50]. However, in another study, out of 15 patients receiving CAR T-cells after previous allogeneic HCT, 3 patients (20%) developed GVHD requiring systemic therapy [30]. This is an important side effect, which might be even more frequent when third-party CAR T cell products are introduced, unless the endogenous T cell receptor is genetically deleted or inactivated. Taken together, GVHD might be a side effect of special interest and if suspected standard therapy is recommended [9].

Subsequent Malignancies
In a study on late effects after CD19 CAR T-cell therapy with a median duration of follow-up of around 28 months, 21 out of 86 patients had ongoing CR. Six of those patients (29%) developed subsequent malignant disease (2 MDS, 1 melanoma, 2 non-melanoma skin cancer, and 1 multiple myeloma) [30]. However, a follow-up period of 2.5 years is most probably not sufficient to estimate the frequency of secondary malignancies after CART, since the development of malignancies is typically associated with latency [51]. Furthermore, it will be difficult to evaluate this risk independently of cytostatic therapy before CART.

Potential Risk of other Long-Term Complications and Implications of the Available Evidence for Clinical Care
Since CAR T-cell therapy is based on genetic manipulation of recipients' lymphocytes, one has to be aware of genotoxicity, such as lymphoma induction/tumor induction, itself [52], To date, this long-term risk cannot be estimated and it is important to perform long-term surveillance. In the future routine checkups for monitoring and prevention, unexpected or rare organ toxicities will be necessary. Data on CAR T cell persistence will also be of major importance to estimate the risk of long-term organ toxicities. Such recommendations will be based on mature outcome data, years after the approval of CART.

Outlook and Future Perspectives
A major future effort will be to reduce CAR T-cell treatment-related toxicities. There are numerous disease-and patient-related variables that contribute to the risk of toxicities, such as co-morbidities, age, tumor type, tumor cell burden, and conditioning therapy. However, modifications in the CAR T-cell product characteristics offer the most promising perspective to decrease toxicities. Specifically, the design of optimized CAR constructs may lead to the opportunity to reduce toxicities. This review is limited by the fact that we mainly focused on two products (Tisa-cel and Axi-cel) in which long-term data/post approval data was available. A variety of new CART are underway, for LBL, such as Liso-cel, mantle cell lymphoma Brexucabtagen autoleucel, and Ide-cel, orva-cel, and the JNJ-4528 for multiple myeloma. Furthermore, a recent report describes the generation of a CD19 41BB construct with a lower production of cytokines and slower proliferation kinetic. No significant elevation in serum cytokine levels after CAR T cell infusion were detected and no CRS or neurotoxicity occurred in the 11 patients treated with this construct [53]. Recently, data on allogeneic CD19 CAR-T cell constructs with an eliminated endogenous T cell receptor were reported. Safety data on this third party off the shelf CART was encouraging, since no GVHD occurred and ≥ • III CRS/ICAN rates were rather low [54].
In addition, a better understanding of the pathophysiology of CAR T-cell-related complications will be critical. Animal models that were developed recently could be the key to a better understanding [2,55]. The development of therapies aimed at the normalization of endothelial dysfunction, which is associated to CAR T-cell-related complications, may be specifically promising [3,23,56].
A major concern regarding CART outside the CD19 setting is on-target off-tumor toxicities. The difficulty is that there is a lack of suitable tumor neo-antigens and the majority antigens that are under investigation in CAR T-cells trials have a considerable expression in normal tissues. As a result, the use of CAR-T cells for the treatment of solid tumors may have an increased risk of severe toxicities as compared with CD19-specific CAR T-cells [57,58].

Conclusions
CD19+ CAR T-cells are promising new treatment options for patients with certain types of leukemia and lymphoma. However, there are considerable toxicities, such as cytokine release syndrome, immune effector cell-associated neurotoxicity syndrome, cardiotoxicity, pulmonary toxicity, metabolic complications, secondary macrophage-activation syndrome, and prolonged cytopenia. Optimization of the management of these unwanted short-term and medium-term effects could further improve clinical outcome.
Funding: This research received no external funding.