CD19-specific chimeric antigen receptor T cell (CAR-T) therapy has generated promising results in children with relapsed/refractory (R/R) B-cell acute lymphoblastic leukemia (ALL), but its use can be complicated by neurotoxicity and severe cytokine release syndrome (sCRS). Patients with central nervous system (CNS) relapse, CNS stage 3 disease, and/or symptomatic neurological disease have frequently been excluded from pediatric studies of CAR-T therapy out of concern for increased neurotoxicity risks.1,2 Data are thus limited regarding the antitumor activity of CAR-T against CNS leukemia (CNSL) or ALL with neurological comorbidities (NC).
The CNS is a primary site of relapse in patients with ALL and is often associated with a less favorable response to therapy.3 Intensive chemotherapy, intrathecal injection (IT), cranial radiation, or even allogeneic hematopoietic stem cell transplantation (allo-HSCT) are the available options for CNS relapse, but these strategies tend to be unsuccessful, especially in patients with multiple relapses.4 Alternative approaches are urgently required, particularly for CNS relapse after cranial radiation or allo-HSCT. A few recent preliminary clinical reports have revealed the feasibility and safety of CAR-T therapy in treating CNSL or bone marrow (BM) relapse with NC.5–8
Neurotoxicity occurs in ~40% of patients, with manifestations ranging from mild delirium to fatal cerebral edema.9 Reported rates of severe neurotoxicity higher than grade 3 are 0–21% in pediatric studies,1,2,10 and neurotoxicity shows a significant association with high pretreatment disease burden, higher peak CAR-T cell expansion, and higher elevations of proinflammatory cytokines in blood.11 CAR-T can effectively eliminate leukemia cells in medullary and extramedullary disease, but a high tumor burden in CNS may cause severe neurotoxicity requiring intense intervention.7 In a large post hoc analysis of CAR-T use in patients with CNS R/R B-ALL, between the CNS-positive and CNS-negative stratum, the incidence and severity of neurotoxicity did not differ (CNS-positive stratum: CNS3 or persistent CNS2 at relapse or within the 12 months preceding infusion).12 However, in multivariate analyses of neurotoxicity in that study, a history of CNS R/R disease and a BM burden of M3 were significantly associated with increased risk. Patients with a BM M3 or NC had a significantly higher risk of developing severe neurotoxicity.
Treatment for R/R ALL with CNSL or neurological complications remains challenging, and guidance for these patients is lacking. Tumor burden reduction by IT or bridging chemotherapy and a return to near-normal neurological status before CAR-T infusion may contribute to reduced sCRS or neurotoxicity, but the persistence of CAR-T cells has not been confirmed.5,7,12 In this study, we reported the ability of humanized CD19 CAR-T cells harboring 4-1BB and the CD3 ξ moiety to induce CR in ALL patients with CNS involvement or NC at an appropriately low disease burden.
MATERIALS AND METHODS
Patients and Ethical Review
We carried out a prospective clinical study of CD19-directed CAR-T cell therapy for R/R ALL from December 2019 to June 2021. The study was registered at chictr.org.cn (No. ChiCTR2000030159) and approved by the Shanghai Children’s Hospital Institutional Review Board. All participants, with the consent of their legal guardians, volunteered to participate in this trial and they and/or their legal guardians provided written informed consent before enrollment. CNS status was classified as CNS1 [cerebrospinal fluid (CSF) with white blood cell (WBC) <5/μL and no blasts], CNS2 (WBC <5/μL and cytology positive for blasts), and CNS3 (WBC ≥5/μL, with the presence of lymphoblasts and/or parenchymal or cranial nerve involvement). The NC-positive subset was defined by an ALL-unrelated neurological complication, including epilepsy, stroke, neurological deficit, and abnormal brain imaging such as posterior reversible encephalopathy syndrome (PRES), leukodystrophy, or encephalatrophy at relapse or within the 12 months preceding CAR-T cell infusion. Patients with NC were controlled for the period of time needed for a return to near-normal neurological status. The patients were followed up to November 30, 2021.
CAR Transduction Procedure
The CAR transgene and CAR-T cell manufacturing were engineered by Shanghai Genbase Biotechnology Co., Ltd. China. Peripheral blood (PB) mononuclear cells were obtained by apheresis. CD3+ T cells were purified by CD3+ microbeads (Miltenyi Biotec, Inc.) and stimulated by anti-CD3/anti-CD28 mAb-coated Human T-Expander beads (cat. no. 11141D; Thermo Fisher Scientific, Inc.) for 24–48 hours. T cells were cultured in T-cell medium X-Vivo 15 (Lonza Group, Ltd.) followed by supplementation with 200 IU/ml rhIL-2 (Proluekin; Novartis International AG). A lentiviral vector encoding CD19 CAR with the humanized signaling domains 4-1BB/CD3ζ (Shanghai Genbase Biotechnology Co., Ltd) was transduced into the activated T cells. The transduction was performed for 48 hours at a multiplicity of infection of 0.3. Transduced T cells were cultured, amplified, and harvested over 5–7 days using the G-Rex system.
Transduction efficiency was defined as the ratio of CAR-T to CD3+ T cells, determined by flow cytometry (FCM) with anti-CD19 CAR-T cell–specific detection reagent. The mean CD19 CAR transduction efficiency was 54.1%±5.8%. The products included both CD4+ (60.5%±14.9%) and CD8+ T cells (33.2%±16.8%) and did not contain any CD19+ B cells. The corresponding subpopulations of naive T cells (CD45RA+CCR7+), central memory T cells (CD45RA−CCR7+), effector memory T cells (CD45RA−CCR7−), and effector T cells (CD45RA+CCR7−) were 57.0%±17.0%, 20.5%±7.1%, 10.2%±6.5%, and 12.3%±6.6%, respectively.
Bridging Chemotherapy and Lymphodepletion Conditioning
For patients with a tumor burden of marrow blasts ≥40%, the bridging regimen included vincristine 1.5 mg/m2 on day 1 and day 8, daunorubicin 30 mg/m2 on day 1 or on day 8 in addition, perasparaginase 2000 IU/m2 on day 3, and dexamethasone 8 mg/m2 on days 1 to 14, with a 7-day taper. Patients with a high CNS burden of (≥5/μL blasts in CSF or solid mass) received IT with cytarabine, methotrexate, and dexamethasone every other day. IT was performed every week or until blasts were <5/μL. The aim was to control CSF blasts <5/μL or deplete parenchymal infiltration with no detection of epileptic waves on electroencephalogram.
Lymphodepletion conditioning was performed with fludarabine 40 mg/m2/day for days −5 to −3 and cyclophosphamide 500 mg/m2/day for days −5 to −4. The day of infusion was defined as day 0, and patients were infused at a single dose of 1×106 cells/kg of CAR-T cells.
Residual blasts were evaluated based on morphology and minimal residual disease (MRD) by FCM and reverse transcription-polymerase chain reaction (if fusion genes were present) in CSF and BM on day 28 after CAR-T infusion. The response for extramedullary invasion was judged by computed tomography (CT), positron emission tomography (PET)-CT, magnetic resonance imaging (MRI), or B ultrasound. Clinical responses were defined as complete remission (CR), CR with incomplete count recovery, partial remission, or no remission (NR) according to the National Comprehensive Cancer Network guidelines. Negativity for MRD was defined as <0.01%, as determined by FCM.
Grading and Management of CRS and Neurotoxicity
CRS and neurotoxicity were graded according to the American Society of Transplantation and Cellular Therapy consensus guidelines.13 Other adverse events were graded using the National Cancer Institute Common Terminology Criteria for Adverse Events, version 4.03. sCRS was defined as grade 3 or higher, and severe neurotoxicity was defined as any seizure or as grade 3 or higher.
Levetiracetam (10 mg/kg, twice daily) was given as anti-epileptic prophylaxis to patients with a history of seizure from the day of CAR-T infusion until 1 month after infusion. Patients already on anti-epileptics continued their home regimen during CAR-T therapy. Seizure prophylaxis with levetiracetam was initiated when patients developed sCRS or new neurological symptoms. Management of immune effector cell–associated neurotoxicity syndrome (ICANS) and CRS was performed based on Neelapu et al,14 with some adaptations. Active seizures were managed with benzodiazepines (eg, diazepam or midazolam) for patients who developed status epilepticus. Mannitol (2.5–5 mL/kg/dose) and furosemide (1-2 mg/kg/dose) were used to control intracranial hypertension. Details of the management of CRS and ICANS are given in the Supplemental Methods, Supplemental Digital Content 1, https://links.lww.com/JIT/A684.
CAR-T Cell Detection
CAR-T cells in PB were measured on the expansion days, on day 14, and monthly from 1 month to the month when CAR-T became undetectable. CAR-T cells in CSF were detected on the peek CRS day or at acute neurotoxicity onset, on day 14, monthly at 1 month to 3 months, and then every 2 or 3 months until CAR-T became undetectable. Cells were incubated for 1 hour with B5338 (anti-CD19 CAR-T cell–specific detection antibody for single-chain Fvs, Genbase Biotechnology, Shanghai, China) as the first antibody, and for 0.5 hours with APC-anti mouse immunoglobulin (Ig) G (BioLegend, San Diego, CA) as the second antibody. The proportion of the CD3 and B5338 double-positive population was distinguished as CAR-T cells. CAR-T cell counts were determined using a FACSCalibur flow cytometer (BD Biosciences) and reported as the fraction of live lymphocyte cells. The absolute number of CAR-T cells was defined as the number per microliter. Details of the detection procedure are given in the Supplemental file p2, Supplemental Digital Content 1, https://links.lww.com/JIT/A684.
Serum and CSF concentrations of target cytokines [interleukin (IL)-1β, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12p70, IL-17, interferon (IFN)-γ, tumor necrosis factor (TNF)-α, and IFN-α] were measured using a multiple microspheres flow immunofluorescence assay (12 cytokine kits, Raisecare, China) according to the manufacturer’s instructions. Serum cytokines were detected within 7 days before CAR-T infusion, on the days of CRS or ICANS onset, and on days 14 and 28. Cytokines in CSF were assessed within 7 days before CAR-T infusion, on the day of peak CRS or ICANS onset, and on days 14 and 28.
B Lymphocytes Detection
PB lymphocyte subsets (CD3+T, CD4+T, CD8+T, NKT, B cells) by percentage and cell count were analyzed with FCM and with Cell Quest software. Antibodies to CD19 (BD Biosciences) were used for B-cell quantification. B-cell recovery was defined as >3% B cells/lymphocytes (the lower limit of normal for blood B-cell levels).15
Non-normally distributed data are presented as medians and full ranges. Comparisons between 2 groups were made using the Mann–Whitney U test or Wilcoxon matched pairs test. Comparisons among 3 or more groups were made using one-way analysis of variance and the Kruskal-Wallis test. Comparisons of two or more factors were made using two-way one-way analysis of variance. Correlations were analyzed by the Spearman correlation, and correlation coefficients were calculated. Data for patients who died without evidence of disease before their first post-infusion assessment were not evaluated. All statistical analyses were performed with GraphPad Prism 5 (GraphPad Software Inc., La Jolla, CA) and SPSS 19.0 (SPSS Inc., Chicago, IL) software. All tests were two-sided, and P<0.05 was considered statistically significant.
A total of 12 pediatric R/R ALL patients were enrolled [4 girls, 8 boys; median age, 7 (range 4–16) years], 2 with primary refractory disease and 10 with a median of 2 (range 1–3) relapses. Four patients had CNSL, 3 had BM relapse with NC, and 5 had BM relapse without NC. Of the 4 patients with CNSL, 3 experienced BM relapses and 1 had bilateral testis relapses. Of the 3 patients with NC, 1 case was complicated with PRES, 1 patient had PRES and epilepsy, and 1 had incomplete myelination accompanied by intrathecal synthesis of IgG.
Two patients who relapsed after allo-HSCT received donor-derived CAR-T. Seven patients received bridging chemotherapy, which failed for 1 of them. All patients with CNSL received IT preceding CAR-T infusion and had CSF blasts <5/μL or obvious depletion of parenchymal infiltration (Fig. 1A–D). One patient with NC and 1 with CNSL received anti-epileptic prophylaxis with levetiracetam (see Table 1 for details).
TABLE 1 -
Patient and Disease Characteristics with CNSL or Neurological Comorbidities
||Age at CAR-T Infusion (mo)
||Cytogenetic/ Molecular Genetic Alterations
||Leukemia Burden Reduction
||BM Blasts Before CAR-T
||CNS Infiltration/Neurological Comorbidities
||BM, CSF, leptomeninges, spleen, bone, muscle
||CNS3, PET-CT abnormal signal
||BM, CSF, cerebral cortex
||CNS3, MRI/ PET-CT abnormal signal
||CNS3, testis infiltration, cranial radiation at 1st testis relapse
||Hyperdiploid, PAX5, IKZF1
||Generalized seizure with PRES on MRI
||TEL/AML1, TP53, NRAS
||Generalized seizure with PRES on MRI, epilepsy, slight encephalatrophy, allo-HSCT
||Death from neurotoxicity
||Incomplete myelination, abnormal oligoclonal bands
||PTPN11, RHOA, FLT3/TKD, t(1;6) (q31;p21)
||Death from relapse
auto indicates autologous-derived; BM, bone marrow; CAR-T, chimeric antigen receptor T cell; CNS, central nervous system; CSF, cerebrospinal fluid; donor, donor-derived; EEG, electroencephalogram; FCM, flow cytometry; haploid, haploidentical transplantation; IT, intrathecal injection; mo, months; MRD, minimal residual disease; MRI, magnetic resonance imaging; MUD, matched unrelated donor; Ph+, Philadelphia chromosome; PRES, posterior reversible encephalopathy syndrome; RTX, rituximab; VDDL, vincristine, daunorubicin, dexamethasone, and perasparaginase.
Eleven patients had a response-evaluable assessment on day 28. The rate of CR was 100%, so that the rate of NR was 0%. One patient (NC2) who died of severe ICANS and sCRS on day 18 had no data available for BM assessment. All 4 patients with CNSL achieved CR (100%) with blast-negative CSF and BM by FCM, with CAR-T cell expansion in CSF and PB (Fig. 1A–D). In 2 patients (CNSL1, CNSL2), MRI or PET-CT was negative for parenchymal infiltration after CAR-T therapy (Fig. 2A–D). The 2 CNSL patients (CNSL1, CNSL4) had other sites of extramedullary infiltration that had become negative on imaging. CR/incomplete count recovery was also 100% among patients with BM relapses, with or without NC, and they all were FCM-MRD negative on day 28.
One patient (NC2) died during CAR-T therapy. He had a CRS of grade 4 and a seizure accompanied with cerebellar edema. This patient ultimately died of severe ICANS on day 18 after CAR-T therapy. Mortality related to CAR-T therapy was 8.3%.
Circulating CAR-T Cells in CSF and PB
CAR-T cells in CSF were measured in 11 of 12 patients (a CSF sample from patient BM1 was unavailable). Circulating CAR-T cells in CSF were found in 9 patients (81.8%) on day 7 (±2) and in all specimens (100%) on day 14. The proportion of CAR-T cells among lymphocytes in the CSF peaked at 39.5% (range 0.2–64.8%; Fig. 3), with peak expansion on day 14 (7–28). The peak CSF CAR-T expansion was higher than in PB, at 39.5% versus 16.2% (P=0.019). CSF CAR-T cells in patients with CNSL and those with BM relapses with and without NC had similar peaks (P=0.899; Fig. 1E); median values were 33.7% (range 9.6–64.8%) with CNSL, 48.2% (range 0.2–49%) with BM relapse and NC, and 34.5% (range 5.3–42.2%) with BM relapse without NC. The peak absolute CAR-T cell count in CSF occurred around day 14 (7–28) with a median of 0.809/μL (range 0.027–6.62/μL). The peak absolute number of CSF CAR-T cells was 2.207 (0.027–6.62)/μL with BM relapse and NC, 0.82 (0.5–2.005)/μL with CNSL, and 0.458 (0.098–2.7)/μL with BM relapse and no NC (P=0.719; Fig. 1F). The median duration of CSF CAR-T cells in patients with CNSL was 5.5 (range 3–9) months, which differed significantly from the 3 (range 2–3) months for patients with isolated BM relapses (P=0.031; Fig. 1G).
The peak fraction of CAR-T cells in PB was similar with CNSL, BM relapse with NC, and BM relapse without NC on peak day 8 (5–11) (P=0.486; Fig. 1E, Supplemental Fig. 1, Supplemental Digital Content 2, https://links.lww.com/JIT/A685). Median values were 209.7 (50–184.4)/μL with CNSL, 72.5 (34.2–170.0)/μL with BM relapse and NC, and 71 (32.8–134.9)/μL with BM relapse and no NC (P=0.233; Fig. 1F). The duration of PB CAR-T cells was similar between CNSL and isolated BM relapses (P=0.339; Fig. 1G). The peak day of CAR-T cells in CSF occurred later than in blood (P=0.005, day 14 vs. day 8), whereas CAR-T cells in CSF remained at a higher fraction than in blood on day 14 (P=0.014) and on day 28 (P=0.016; Fig. 1I). The peak for PB CAR-T cells in fractional and absolute counts was higher in patients with MRD+/− than for BM with residual blasts (P=0.028, P=0.202; Fig. 1H). The persistence of CAR-T cells was similar between 2 groups with medians of 3 (2–9) months with MRD+/– and 4 (1–4) months with BM with residual blasts (P=0.917).
Neurotoxicity developed in 2 patients. Patient NC2 had grade 5 ICANS, and patient CNSL2 had chronic neurotoxicity with white matter lesions at 1 month after CAR-T infusion. The 2 patients who experienced neurotoxicity did not have greater peak CAR-T cell values in CSF (0.2% and 9.6%, and 0.027 and 0.809 cells/μL, respectively) compared with those without neurotoxicity (40.5%, range 5.3–64.8%; 0.839 cells/μL, range 0.098–6.62 cells/μL).
Neurotoxicity in CNSL and Neurological Comorbidities
Initial MRI or CT/PET-CT imaging of CNS infiltration was obtained before lymphodepletion chemotherapy. Patients CNSL1 and CNSL2 had abnormal PET-CT brain imaging (Fig. 2A, B). CNSL2 had extensive leptomeninges infiltration with hyperintensities on enhanced T2-FLAIR images (Fig. 2C) at 2 months before CAR-T cell infusion. Patient NC2 had a history of epilepsy and PRES, and T1-weighted images at 1 month before CAR-T cell infusion showed slight encephalatrophy and ventriculomegaly (Fig. 2G). Electroencephalogram monitoring showed abnormal waveforms in all 3 of these patients (CNSL1, CNSL2, and NC2), 2 of whom (CNSL2 and NC2) developed neurotoxicity. Patient CNSL2 did not have neurological symptoms at CRS peak, but the brain MRI showed leukodystrophy on day 30 (Fig. 2E). Routine CSF testing was normal in this patient, with WBC 0/μL, glucose 2.7 mmol/L, and protein 200 mg/L. IgM and IgG for toxoplasma, rubella virus, cytomegalovirus, and herpes simplex virus (HSV) in CSF were negative. The DNA of HHV-6, HSV-1, HSV-2, and JC virus in CSF was not detectable. The patient received IgG 1 g/kg for 2 days and gradually recovered until 1 year post CAR-T infusion (Fig. 2F). Patient NC2 had a seizure with a decreased level of consciousness and required invasive ventilatory support on day 13. Brain CT showed acute cerebellar edema and subsequently cerebral hemorrhage coincident with thrombocytopenia (Fig. 2H). This patient ultimately died of ICANS 5. None of the 5 patients without NC developed neurotoxicity (Fig. 2I).
Cytokines in CNSL and Neurological Comorbidities
In the CSF, IL-6, IL-8, and IL-10 had risen significantly by the peak of CRS or ICANS onset (P=0.009, 0.005, 0.001, respectively), from a median of 6.1 (2–42.4) pg/mL to 124.2 (15.4–5618) pg/mL for IL-6, 100 (21.9–205.8) pg/mL to 342.3 (70.4–1755) pg/mL for IL-8, and 2 (1.6–2) pg/mL to 17.3 (2–56.1) pg/mL for IL-10 in the whole cohort. Other cytokines, including IL-1β, IL-2, IL-5, and IFN-γ, showed mild but significant elevations (P=0.016, 0.004, 0.031, 0.028, respectively), although median values remained in the normal range (see Table 2 for details). At the acute phase of CRS, CSF IL-6, IL-8, and IL-10 showed extreme increase in patients with NC than those with CNSL and no NC (for comparison of each biomarker among the 3 groups: IL-6, P=0.041; IL-8, P=0.016; IL-10, P=0.177; Fig. 4A–C). IL-2 was only mildly elevated overall, whereas IL-2 increased significantly in patients with NC (P=0.010; Fig. 4G). Results for other biomarkers are shown in Supplemental Fig. 2, Supplemental Digital Content 3, https://links.lww.com/JIT/A686.
TABLE 2 -
CSF and Serum Cytokines from Baseline to Peak CRS After CAR-T Therapy
|CSF (pg/mL) Median (range)
|Serum (pg/mL) Median (range)
P-value indicates the comparison with peak and baseline.
CAR-T indicates chimeric antigen receptor T cell; CRS, cytokine release syndrome; CSF, cerebrospinal fluid; IFN, interferon; IL, interleukin.
In serum, IL-6, IL-2, IL-10, IFN-γ, IL-5, and IL-1β had increased significantly by the peak phase (P=0.001, 0.003, 0.001, 0.001, 0.006, 0.008, respectively; Table 2 for details). IL-8 and IL-4 rose slightly but remained within normal range ( P=0.005, 0.047, respectively; Table 2). IL-12P70, IL-17, IFN-α, and TNF-α did not increase (P=0.063, 0.063, 0.063, 0.078, respectively). Cytokine release in serum after CAR-T therapy was seen in Supplemental Fig. 3, Supplemental Digital Content 4, https://links.lww.com/JIT/A687.
Cytokine Comparisons between Compartments
In comparing biomarker changes between compartments, we found that among patients with NC, in the paired samples, IL-6 levels were higher in CSF (median 847.85 pg/mL) than in serum (66.6 pg/mL) (Fig. 4D). In patients with CNSL or no NC, IL-6 was lower in CSF than in the serum (CNSL: median 19.8 pg/mL CSF vs. 33.6 pg/mL serum; without NC: 102.9 pg/mL CSF vs. 493.9 pg/mL serum; Fig. 4D). IL-8 levels were higher in CSF than in PB before CAR-T and at peak CRS (P=0.027, 0.0002, respectively; Fig. 4E), but IL-10 did not differ in the paired samples at peak CRS (P=0.547; Fig. 4F).
Cytokines and Neurotoxicity
When we examined biomarkers in each compartment in association with neurotoxicity, we found that IL-6, IL-8, and IL-2 in CSF and IL-6, IL-8, IL-2, IL-5, IL-10, and IFN-γ in PB were markedly elevated (over 10 times normal levels) during acute neurotoxicity (patient NC2). IL-1β and IL-5 in CSF and IL-1β in PB were only mildly elevated; however, IL-10 and IFN-γ remained at normal levels in CSF. IL-4, IL-12P70, IL-17, IFN-α, and TNF-α in PB and CSF remained normal during acute neurotoxicity. Further analysis showed no correlation between IL-6 elevation and CSF CAR-T cell expansion (P=0.126).
WBCs and Proteins in the CSF
CSF WBCs had increased by the CRS peak from a median of 1 (0–7) cells/μL to 10 (1–46) cells/μL (P=0.001; Fig. 4H) in the whole group, and from 3 (0–7) cells/μL to 8 (3–20) cells/μL with CNSL (P=0.097), from 1 (0-1) cells/μL to 19 (2–46) cells/μL with NC (P=0.250), and from 1 (0–2) cells/μL to 12 (1–23) cells/μL without NC (P=0.098). CSF WBCs did not differ among the 3 patient subsets (P=0.752). CSF WBCs in the patient with ICANS remained normal at 2 cells/μL. CSF protein rose from a median of 235 (range 210–940) mg/L to 330 (range 240–1070) mg/L in the whole group (P=0.060) and from 230 (210–250) mg/L to 380 (260–910) mg/L in those with BM relapses (P=0.008; Fig. 4I). CSF protein in patients with CNSL decreased from 315 (210–940) mg/L to 235 (220–780) mg/L (P=0.250) and increased most significantly in the patient with ICANS, from 210 mg/L to 910 mg/L.
Eleven of 12 patients survived CAR-T–related toxicity, and 1 patient (NC2) died of ICANS. One patient (BM4) experienced relapse in the second month after CD19 CAR-T and received CD19/22 CAR-T therapy in the third month. The disease did not respond positively to remission, and the patient died in NR (Fig. 2J). Ten patients remained in CR, 8 of whom underwent allo-HSCT and 2 of whom had CNSL and did not undergo HSCT. One-year event-free survival was 90.9%±8.7%, with a median of 12 (0–22) months of follow-up.
Currently, treatment for R/R ALL remains a difficult challenge, especially with CNS relapse or neurological complications. The most common site of extramedullary relapse is the CNS, representing ~20% of the total.16,17 For patients with CNSL that is refractory to CNS-directed therapies, few treatment options are available.18 CAR-T cells can migrate into a leukemia cell sanctuary such as the CSF by overcoming the blood–brain barrier (BBB), which may provide an opportunity for patients whose CNSL is refractory to conventional therapeutic regimens.19
Nevertheless, neurotoxicity remains a severe and potentially fatal event with CAR-T therapy. Most clinical trials of CAR-T have excluded patients with CNSL or symptomatic neurological disease because of a potentially increased risk of neurotoxicity.1,2,20 For the wider application of CAR-T therapy, patients with CNS recurrences have been intermittently enrolled to explore the feasibility of eliminating the blasts in the CNS.8,21 However, the case numbers are limited for pediatric patients. In this prospective cohort study of children undergoing CD19-directed CAR-T cell treatment, we performed comprehensive CSF analyses, CAR-T cell detection, cytokines quantification, and imaging scans, which together may provide new insights into R/R ALL with CNSL or NC and the use of CAR-T cell therapy for these patients.
Humanized CAR-T cells showed good efficacy in inducing remission in the CNS in our cohort. High rates of detectable CAR-T cells in the CSF offer evidence that CAR-T cells can traffic to extramedullary sites. In our data, CAR-T cells persisted in the CSF of patients with CNSL for a median of 5.5 months, with the longest durability of 9 months. Results from a published clinical trial showed that CAR-T cells with a 4-1BB domain persisted for more than 6 months in most cases.22 However, the durability of our native CAR-T products in CSF was short in those with BM recurrences, with a median of 3 months. We found that patients with CNSL had a longer duration of CAR-T cells in the CSF than those without CNS infiltration. It should be noted that the patients with CNSL in our cohort had either BM relapse or testis recurrence. He et al21 found that patients with only CNS recurrence had higher levels of CAR-T cells in the CSF and relatively lower severity of toxic effects than those with BM and CNS recurrence. Li et al23 observed that the peak fraction of CAR-T cells in the CSF during recurrent CNSL was higher compared with PB, consistent with our results. We further confirmed that the peak fraction of CSF CAR-T cells with isolated BM relapses was also higher in CSF than in PB. We performed bridging chemotherapy in patients with high tumor burden, but a lower tumor burden did not appear to reduce the peak level or persistence of CAR-T cells.
ICANS has previously been reported in 40% of children and young adults with ALL (13% severe) and 50% of adults (50% severe).10,24 Risk predictors for developing neurotoxicity are not clear, but factors that have been considered include higher disease burden, higher CAR-T cell expansion, higher proinflammatory cytokines, and extramedullary disease.11,25,26 To date, the impact of CNS disease burden on ICANS is not well defined.11,25 In a study of patients who retained >5/μL blasts in CSF or had a solid mass before CAR-T cell expansion, 80% developed severe neurotoxicity that featured persistent cerebral edema and seizure,7 even though reducing a high tumor burden in CNS was expected to decrease neurotoxicity severity. Rubinstein et al27 reported that 7 pediatric patients with controlled NC or CNSL under CNS2 were infused with tisagenlecleucel and experienced mild to no neurotoxicity. One report described 3 adult patients with CNSL at stage CNS3 (blasts >5/μL) who had undergone burden reduction (blasts ≤5/μL) with chemotherapy before CAR-T cell infusion. They all achieved CR accompanied by mild neurotoxicity.21 In the post hoc analysis of CAR-T use, CNS2 or CNS3 at infusion was not associated with a higher risk for any or severe neurotoxicity, although patients in the R/R CNSL stratum were more likely to develop any symptoms of neurotoxicity.12
In addition, a strong association between BM burden and the risk of CRS and neurotoxicity has been reported.12,24 Because of the small number of patients in our study, we could not analyze correlations between BM burden and neurotoxicity risk. Neurotoxicity might be mitigated by a low burden in BM disease.12 Our choice for reducing tumor burden was bridging chemotherapy or enhanced lymphodepletion, which remain controversial among clinicians. We conducted IT and/or intensive chemotherapy to reduce the potential risk of severe neurotoxicity, as well as symptomatic NC controlled before CAR-T infusion. In our cohort, one of 7 patients in the group with high tumor burden did not experience blast reduction to under 20%, whereas in all patients with CNSL, blasts were controlled to <5/μL, and parenchymal lesions were controlled, as well. CRS of grade 1-2 and no ICANS were seen in patients with well-controlled tumor burden. Life-threatening cerebral edema occurred in a patient who did not experience blast depletion.
The increased permeability of the BBB and increased cytokines caused by the activation of CNS endothelial cells contribute to neurotoxicity in CAR-T therapy.25 Immune effector cells can pass into the CNS, where they can have a localized inflammatory function. In fact, proinflammatory cytokines in the CSF likely accumulate both from the circulating blood and from CNS production.11,25 CAR-T cells were present in all CSF specimens from our patients, similar to previously reported findings.28,29 Not all inflammatory cytokines were significantly elevated with CAR-T cell expansion in the CSF, however, suggesting that infiltrating CAR-T cells do not have a primary role in the development of ICANS. In our cohort, patients received a fixed dose of CAR-T cells, but the cells themselves vary qualitatively among patients, which may influence cytokine production. We speculated that the toxicity is primarily mediated by the inflammatory cytokine surge that accompanies CAR-T cell expansion, rather than by the CAR-T cells themselves. The role of direct infiltration by CAR-T cells as a driver of neurotoxicity is still unclear. Necropsy from a nonhuman primate monkey model showed brain intraparenchymal infiltration of CAR-T cells,30 but CAR-T cells were not seen on autopsy of a patient who died from cerebral edema.31 It is possible that preautopsy lysis of lymphocytes followed steroid treatment.
In our data, the key inflammatory cytokines IL-6, IL-8, and IL-10 were significantly elevated in CSF after CAR-T therapy and comparably so between patients with CNSL and BM recurrences. These findings imply that local cytokine production in patients with CNSL with a low disease burden is not a prominent feature of neurotoxicity. Cytokines were enriched in CSF in the patient experiencing acute neurotoxicity, with disproportionately high levels of IL-6, IL-8, and IL-2, and mild elevation in IL-1β and IL-5, even as CSF CAR-T cells remained at low concentrations. Limited CSF data are available from CAR-T recipients, but IL-6 expression has been reported to be related to endothelial activation and dysfunction of the BBB, which is the key mediator of CRS and modulates CNS response to injury.11,25,32 IL-1 is reported to have a direct role in ICANS in a humanized mouse model,33 and IL-1α in CSF11 and in serum32 has been correlated with ICANS severity. IL-1β also has been identified as a key mediator of CNS injury, inflammation, and neurodegeneration, although its association with ICANS remains indeterminate.32,33 IL-8 and IL-10 are increased in ICANS, but a causative role for these cytokines has not been established.11,25,28 IL-2, IL-4, IL-17a, IL-15, IFN-α, and IFN-γ all have been reported to occur at high levels in the CSF with CNS involvement,11,25,34,35 which is not entirely consistent with our results. This divergence may be explained by different CAR constructs leading to release of different cytokines.
Previous studies have shown that patients with NC, such as seizure, stroke, brain imaging abnormalities, or neurological deficit, may have an increased risk for neurotoxicity under CAR-T therapy.12,25,28 For the paired specimens of blood and CSF that we evaluated in NC patients, IL-6 increased more in CSF than in blood. CSF IL-6, IL-2, and IL-8 in these patients occurred at much higher levels than in CNSL patients and patients without NC, suggesting increased permeability of the BBB in NC. CSF protein levels increased in our cohort at CRS onset, especially among those with NC. The increased CSF protein might indicate BBB dysfunction after CAR-T infusion.11 We also observed that CSF protein in patients with CNSL decreased after CAR-T infusion, indicating depletion of blasts in the CNS. Although ICANS is reversible in most patients, cerebral edema remains a severe event in clinical trials.36 Brain MRI may be normal or show vasogenic cerebral edema, leptomeningeal enhancement or microhemorrhages with neurotoxicity.37 Long-term toxicity following CAR-T in reports is relatively low.27 In the current study, 1 patient with CNSL experienced chronic toxicity with white matter lesions that gradually recovered after 1 year, suggesting that cellular therapy approaches might offer a tolerable treatment for this case presentation.
In summary, we describe outcomes with CD19 CAR-T cell therapy in 7 cases of pediatric ALL with CNSL and/or NC, and in 5 cases with BM relapse and no NC. We reduced the tumor burden before the initiation of CAR-T therapy, showing a potentially effective method for preventing life-threatening ICANS. The results demonstrated that CAR-T therapy is a safe and effective treatment option for clearance of extramedullary infiltration and medullary blasts in patients with neurological complications, if the tumor burden or neurological disease can be adequately controlled. Further and larger studies are needed to verify these findings. The clinical features and biological correlates of neurotoxicity need to be deeply explored to mitigate toxicity.
CONFLICTS OF INTEREST/FINANCIAL DISCLOSURES
Supported by grants from Shanghai Children’s Hospital (No. 2020ylym09) and Shanghai Shen Kang Hospital Development Center (SHDC22020210). All authors have declared that there are no financial conflicts of interest with regard to this work.
1. Lee DW, Kochenderfer JN, Stetler-Stevenson M, et al. T cells expressing CD19 chimeric antigen receptors for acute lymphoblastic leukaemia in children and young adults: a phase 1 dose-escalation trial. Lancet. 2015;385:517–528.
2. Maude SL, Laetsch TW, Buechner J, et al. Tisagenlecleucel in children and young adults with B-cell lymphoblastic leukemia. N Engl J Med. 2018;378:439–448.
3. Krishnan S, Wade R, Moorman AV, et al. Temporal changes in the incidence and pattern of central nervous system relapses in children with acute lymphoblastic leukaemia treated on four consecutive Medical Research Council trials, 1985-2001. Leukemia. 2010;24:450–459.
4. Sun W, Malvar J, Sposto R, Verma A, et al. Outcome of children with multiply relapsed B-cell acute lymphoblastic leukemia: a therapeutic advances in childhood leukemia & lymphoma study. Leukemia. 2018;32:2316–2325.
5. Rubinstein JD, Nelson AS, Krupski C, et al. Chimeric antigen receptor T-cell therapy in patients with neurologic comorbidities. Pediatr Blood Cancer. 2020;67:e28199.
6. Zhang L, Zuo Y, Lu A, et al. Safety and Efficacy of Chimeric Antigen Receptor T-Cell Therapy in Children with Central Nervous System Leukemia. Clin Lymphoma Myeloma Leuk. 2021;21:e410–e414.
7. Tan Y, Pan J, Deng B, et al. Toxicity and effectiveness of CD19 CAR T therapy in children with high-burden central nervous system refractory B-ALL. Cancer Immunol Immunother. 2021;70:1979–1993.
8. Dai H, Zhang W, Li X, et al. Tolerance and efficacy of autologous or donor-derived T cells expressing CD19 chimeric antigen receptors in adult B-ALL with extramedullary leukemia. Oncoimmunology. 2015;4:e1027469.
9. Gust J, Taraseviciute A, Turtle CJ. Neurotoxicity associated with CD19-targeted CAR-T cell therapies. CNS Drugs. 2018;32:1091–1101.
10. Gardner RA, Finney O, Annesley C, et al. Intent-to-treat leukemia remission by CD19 CAR T cells of defined formulation and dose in children and young adults. Blood. 2017;129:3322–3331.
11. Santomasso BD, Park JH, Salloum D, et al. Clinical and biological correlates of neurotoxicity associated with CAR T-cell therapy in patients with B-cell acute lymphoblastic leukemia. Cancer Discov. 2018;8:958–971.
12. Leahy AB, Newman H, Li Y, Liu H, et al. CD19-targeted chimeric antigen receptor T-cell therapy for CNS relapsed or refractory acute lymphocytic leukaemia: a post-hoc analysis of pooled data from five clinical trials. Lancet Haematol. 2021;8:e711–e722.
13. Lee DW, Santomasso BD, Locke FL, et al. ASTCT consensus grading for cytokine release syndrome and neurologic toxicity associated with immune effector cells. Biol Blood Marrow Transplant. 2019;25:625–638.
14. Neelapu SS, Tummala S, Kebriaei P, et al. Chimeric antigen receptor T-cell therapy- assessment and management of toxicities. Nat Rev Clin Oncol. 2018;15:47–62.
15. Mueller KT, Waldron E, Grupp SA, et al. Clinical pharmacology of tisagenlecleucel in B-cell acute lymphoblastic leukemia. Clin Cancer Res. 2018;24:6175–6184.
16. Nguyen K, Devidas M, Cheng SC, et al. Factors influencing survival after relapse from acute lymphoblastic leukemia: a Children’s oncology group study. Leukemia. 2008;22:2142–2150.
17. Cooper SL, Brown PA. Treatment of pediatric acute lymphoblastic leukemia. Pediatr Clin North Am. 2015;62:61–73.
18. Pan J, Yang JF, Deng BP, et al. High efficacy and safety of low-dose CD19-directed CAR-T cell therapy in 51 refractory or relapsed B acute lymphoblastic leukemia patients. Leukemia. 2017;31:2587–2593.
19. Grupp SA, Kalos M, Barrett D, et al. Chimeric antigen receptor-modified T cells for acute lymphoid leukemia. N Engl J Med. 2013;368:1509–1518.
20. Karschnia P, Jordan JT, Forst DA, et al. Clinical presentation, management, and biomarkers of neurotoxicity after adoptive immunotherapy with CAR T cells. Blood. 2019;133:2212–2221.
21. He X, Xiao X, Li Q, et al. Anti-CD19 CAR-T as a feasible and safe treatment against central nervous system leukemia after intrathecal chemotherapy in adults with relapsed or refractory B-ALL. Leukemia. 2019;33:2102–2104.
22. Maude SL, Frey N, Shaw PA, et al. Chimeric antigen receptor T cells for sustained remissions in leukemia. N Engl J Med. 2014;371:1507–1517.
23. Li X, Liu MJ, Mou N, et al. Efficacy and safety of humanized CD19 CAR-T as a salvage therapy for recurrent CNSL of B-ALL following murine CD19 CAR-T cell therapy. Oncol Lett. 2021;22:788.
24. Turtle CJ, Hanafi LA, Berger C, et al. CD19 CAR-T cells of defined CD4+:CD8+ composition in adult B cell ALL patients. J Clin Invest. 2016;126:2123–2138.
25. Gust J, Hay KA, Hanafi LA, et al. Endothelial activation and blood-brain barrier disruption in neurotoxicity after adoptive immunotherapy with CD19 CAR-T Cells. Cancer Discov. 2017;7:1404–1419.
26. Park JH, Rivière I, Gonen M, et al. Long-term follow-up of CD19 CAR therapy in acute lymphoblastic leukemia. N Engl J Med. 2018;378:449–459.
27. Rubinstein JD, Krupski C, Nelson AS, et al. Chimeric antigen receptor T cell therapy in patients with multiply relapsed or refractory extramedullary leukemia. Biol Blood Marrow Transplant. 2020;26:e280–e285.
28. Gust J, Finney OC, Li D, et al. Glial injury in neurotoxicity after pediatric CD19-directed chimeric antigen receptor T cell therapy. Ann Neurol. 2019;86:42–54.
29. Rheingold SR, Chen LN, Maude SL, et al. Efficient trafficking of chimeric antigen receptor (CAR)-modified T cells to CSF and induction of durable CNS remissions in children with CNS/combined relapsed/refractory ALL. Blood. 2015;126:3769.
30. Taraseviciute A, Tkachev V, Ponce R, et al. Chimeric antigen receptor T cell-mediated neurotoxicity in nonhuman primates. Cancer Discov. 2018;8:750–763.
31. Torre M, Solomon IH, Sutherland CL, et al. Neuropathology of a case with fatal CAR T-cell-associated cerebral edema. J Neuropathol Exp Neurol. 2018;77:877–882.
32. Wehrli M, Gallagher K, Chen YB, et al. Single-center experience using anakinra for steroid-refractory immune effector cell-associated neurotoxicity syndrome (ICANS). J Immunother Cancer. 2022;10:e003847.
33. Gust J, Ponce R, Liles WC, et al. Cytokines in CAR T cell-associated neurotoxicity. Front Immunol. 2020;11:577027.
34. Hu K, Wang Y, Teng X, et al. Cell subsets and cytokine dynamics in cerebrospinal fluid after CAR-T cell therapy for B-cell acute lymphoblastic leukemia with central nervous system involvement. Bone Marrow Transplant. 2021;56:3088–3090.
35. Upreti D, Bakhshinyan D, Bloemberg D, et al. Strategies to enhance the efficacy of T-cell therapy for central nervous system tumors. Front Immunol. 2020;11:599253.
36. Senior M. CAR-T death strikes Kite. Nat Biotechnol. 2017;35:492.
37. Maude SL, Barrett D, Teachey DT, et al. Managing cytokine release syndrome associated with novel T cell-engaging therapies. Cancer J. 2014;20:119–122.