Table of Contents
- Key Points
- Background: Why This Research Matters
- What Is GD2-CART01?
- How the Study Was Conducted
- Who Took Part in the Study
- Making the Treatment for Each Patient
- Safety Results: Side Effects and Toxic Effects
- The Safety Switch: How the Suicide Gene Worked
- How the CAR T Cells Grew and Survived in the Body
- Treatment Responses: How Well It Worked
- Survival Results
- Three Patients With an Unusual and Encouraging Course
- What These Findings Mean for Patients
- Limitations of the Study
- What Patients and Families Should Know
- Frequently Asked Questions
- Source Information
Key Points
- In a 27-patient trial, GD2-CART01 was feasible and safe for children with relapsed or refractory high-risk neuroblastoma, with no dose-limiting toxic effects.
- Cytokine release syndrome occurred in 74% of patients but was mild in 95% of those cases; no central neurotoxic effects developed.
- Seventeen of 27 patients (63%) responded, including 9 complete and 8 partial responses; at the recommended dose, 3-year overall survival was 60%.
- The built-in safety switch worked in one patient, rapidly eliminating transferred cells; the cells later re-expanded and the patient remained in complete response.
- This was a small, single-center trial without a comparison group, so GD2-CART01 remains investigational and is not yet a standard treatment option.
Background: Why This Research Matters
Chimeric antigen receptor (CAR) T-cell therapy is a valuable treatment for relapsed or refractory B-cell cancers. In solid tumors, however, the path to an effective CAR T-cell therapy has been more difficult.
Neuroblastoma is the most common extracranial solid tumor in children. It is responsible for 11% of all deaths from cancer in the pediatric population. Almost half of patients have high-risk disease at diagnosis, and the 5-year event-free survival among these patients is 40 to 50%.
Children in whom first-line therapy fails have a very low probability of recovery with subsequent treatments. Their prognosis is dismal, with long-term survival of approximately 5 to 10%.
Neuroblastoma cells express high levels of a molecule called disialoganglioside GD2. Targeting GD2 with monoclonal antibodies has been associated with a significant increase in survival among high-risk patients. These findings indicate that neuroblastoma is sensitive to immunotherapy and that GD2 is a relevant target.
Few early-phase clinical trials of GD2-targeted CAR T cells had shown that the approach was feasible, and few objective responses had been reported. The researchers developed a GD2-directed CAR construct that incorporates two costimulatory domains: CD28 and 4-1BB.
To address potential neurotoxic effects (harm to the nervous system) associated with third-generation CAR T cells, the researchers also included the gene for inducible caspase 9 (iC9) as a safety switch. This switch allows the transferred cells to be killed if they cause dangerous toxic effects. This article reports the 3-year results of the phase 1–2 clinical trial.
What Is GD2-CART01?
GD2-CART01 is an autologous CAR T-cell therapy. "Autologous" means the cells come from the patient's own body. "Third-generation" means the CAR construct includes two costimulatory domains, CD28 and 4-1BB, which help the T cells activate and persist.
The treatment works by reprogramming a patient's own T cells to recognize and attack cells that carry the GD2 molecule. Neuroblastoma cells carry high levels of GD2, which makes them a target for this therapy.
The therapy also contains an inducible caspase 9 (iC9) suicide gene. This gene acts as a safety switch. If the transferred cells cause dangerous side effects, doctors can activate the switch to eliminate them.
How the Study Was Conducted
The trial was an academic, phase 1–2 clinical trial conducted at the IRCCS Ospedale Pediatrico Bambino Gesù in Rome. The trial assessed the feasibility and safety of GD2-CART01 in patients with relapsed or refractory high-risk neuroblastoma. It also tested the treatment efficacy of the recommended dose.
The phase 1 portion was a dose-finding trial. It followed an escalation and de-escalation schema. The first patients received GD2-CART01 at a dose of 3×10⁶ CAR-positive T cells per kilogram of body weight. Two dose-escalation levels were specified: 6 and 10×10⁶ CAR-positive T cells per kilogram.
If the initial dose level was associated with dose-limiting toxic effects, two additional de-escalation dose levels were planned: 2 and 1×10⁶ CAR-positive T cells per kilogram. Both the national competent authority of Italy and the institutional review board of IRCCS Ospedale Pediatrico Bambino Gesù approved the trial.
Patients between the ages of 1 and 25 years with a diagnosis of high-risk neuroblastoma were enrolled in the phase 1 trial if they had relapsed disease. Patients were also enrolled if they had persistent or progressive disease during first-line treatment.
In the phase 2 trial, patients with metastatic disease and MYCN amplification could be enrolled after completion of first-line treatment, even if they had a complete response. MYCN is a gene; MYCN amplification means the gene is present in extra copies.
Before infusion of GD2-CART01, all patients received the same lymphodepleting chemotherapy. This chemotherapy uses fludarabine and cyclophosphamide. Lymphodepletion means reducing the number of lymphocytes (a type of immune cell) to make room for the new CAR T cells.
Disease response was evaluated with the Revised International Neuroblastoma Response Criteria, using European International Society of Paediatric Oncology Neuroblastoma Group (SIOPEN) scoring. The first disease assessment was performed in every patient at week 6 after GD2-CART01 infusion.
Radiologic and magnetic resonance images and SIOPEN scores were evaluated in a blinded fashion. A dedicated radiologist and a nuclear radiologist performed this review. High disease burden was defined as at least one of the following:
- Bulky disease, defined as a mass that was at least 5 cm in the greatest dimension
- Bone marrow infiltration of greater than 50%
- An MIBG (metaiodobenzylguanidine) SIOPEN score higher than 7, on a scale of 0 to 72, with higher scores indicating greater dissemination of disease
For the statistical analysis, patient characteristics and trial outcomes were described with usual summary statistics. These included counts and percentages for categorical factors, and the mean, median, or range for continuous factors. The median follow-up was derived with Schemper's method.
Cox models with one or two covariates were used to derive hazard ratios with 95% confidence intervals for overall and event-free survival. The assumption of proportional hazards was assessed with the use of supremum tests.
Who Took Part in the Study
A total of 30 patients underwent screening, and 27 were enrolled and received treatment between January 2018 and October 2021. Three patients were excluded: 2 because of rapidly progressing disease and 1 because of a psychiatric disorder.
Of the 27 patients, 18 (67%) were male and 9 (33%) were female. The median age was 6.7 years, with a range of 2.7 to 18.6 years. The median number of previous treatments was 3, with a range of 1 to 6.
At enrollment, 12 patients (44%) had refractory disease, 14 patients (52%) had relapsed disease, and 1 patient (4%) had no evidence of disease after NB-HR01 first-line treatment. NB-HR01 first-line treatment consisted of therapy according to the SIOPEN Trial for High-Risk Neuroblastoma protocol. That therapy included an induction phase with intensive chemotherapy, surgery, high-dose chemotherapy with reinfusion of autologous stem cells, radiotherapy, differentiating treatment with isotretinoin, and immunotherapy with the anti-GD2 monoclonal antibody.
Fourteen patients (52%) had previously received treatment with an anti-GD2 monoclonal antibody. Thirteen patients had not received previous treatment with the GD2-directed monoclonal antibody because they did not have approved indications for its use.
For MYCN status, 7 patients (26%) had amplification, 5 (19%) had gain, 10 (37%) had normal status, and 5 (19%) had unknown status.
For sites of disease involvement:
- Bone: 21 patients (78%)
- Bone marrow: 12 patients (44%)
- Lymph nodes: 11 patients (41%)
- Abdomen: 4 patients (15%)
- Paravertebral area: 7 patients (26%)
- Thorax or pleura: 2 patients (7%)
- Liver: 1 patient (4%)
Before infusion, 18 patients (67%) had an MIBG score of 7 or lower, and 9 patients (33%) had an MIBG score higher than 7. Before enrollment, all patients had disease resistance to at least two lines of treatment, with a range of two to six lines.
One patient in the phase 2 trial had received GD2-CART01 during complete remission at the end of first-line therapy. Of the 26 patients with relapsed or refractory disease, 24 had active, stage 4, metastatic disease at screening before treatment. Six of these patients had bulky disease.
Making the Treatment for Each Patient
GD2-CART01 was manufactured successfully in all patients. No failure to generate GD2-CART01 was observed.
The median total number of CAR-positive T cells at the end of production was 2×10⁹ ± 1.4×10⁹ cells per kilogram, with a range of 0.7 to 6.9×10⁹. The median viability was 90.5 ± 3.7%, with a range of 82.2 to 96.1%. The median transduction efficiency was 72.4 ± 9.9%, with a range of 50.8 to 87.3%.
Transduction efficiency means the percentage of cells that successfully received the new genetic instructions. Because production yielded multiple doses of cells for each patient, 11 patients received multiple infusions. Each patient could receive up to four infusions.
The retrovirus encoding the construct iCasp9.2A.GD2.CD28.4-1BB.zeta is approved for clinical use and has been described previously.
Safety Results: Side Effects and Toxic Effects
No dose-limiting toxic effects were reported in the phase 1 trial. The recommended dose of 10×10⁶ CAR-positive T cells per kilogram was determined for the phase 2 portion of the trial.
Cytokine release syndrome (CRS) was one of the most common drug-related adverse events. CRS is a condition in which immune cells release large amounts of inflammatory chemicals into the blood. CRS was observed in 20 of 27 patients (74%) after the first infusion.
These CRS events were grade 1 or 2 in 19 of 20 patients (95%), meaning they were mild. Grade 3 CRS developed in only 1 patient after the first infusion, and it rapidly resolved after administration of tocilizumab. Tocilizumab is a drug that blocks the inflammatory signal interleukin-6.
Serum levels of interferon-γ, interleukin-6, tumor necrosis factor α, and interleukin-10 were monitored in all patients. The researchers observed a correlation between significantly higher levels of circulating cytokines and the development and grade of cytokine release syndrome.
Central neurotoxic effects did not develop in any of the patients. Peripheral neurotoxic effects or pain occurred in 6 patients, all grade 1 or 2.
Severe hematologic toxic effects (effects on blood cells) developed in all patients. These were induced by lymphodepletion and were sustained after GD2-CART01 infusion. The hematologic toxic effects included:
- Anemia: grade 1 or 2 in 8 patients, grade 3 in 19 patients
- Neutropenia: grade 4 in all 27 patients
- Thrombocytopenia: grade 1 or 2 in 1 patient, grade 3 in 4 patients, grade 4 in 19 patients
For abnormal laboratory values, elevated alanine aminotransferase level, aspartate aminotransferase level, or both occurred in 13 patients at grade 1 or 2 and in 7 patients at grade 3. Elevated bilirubin level occurred in 3 patients at grade 1 or 2 and in 1 patient at grade 3. Electrolyte abnormalities occurred in 4 patients at grade 1 or 2 and in 2 patients at grade 3.
For miscellaneous events, a Clostridium difficile infection occurred in 1 patient at grade 3. Rash occurred in 3 patients at grade 1 or 2. Dysuria (painful urination) occurred in 2 patients at grade 1 or 2. One patient had a grade 4 brain hemorrhage.
Grade 3, transient hepatic toxic effects (liver effects) developed in seven patients after infusion. In five of these patients, hepatic toxic effects were present before enrollment or before GD2-CART01 infusion and transiently worsened with the development of cytokine release syndrome. In one of the other two patients, a clinically relevant hepatic infiltration of disease was present and led to grade 3 toxic effects on activation of GD2-CART01. The other patient had grade 3 hepatic toxic effects 3 weeks after infusion of GD2-CART01, after resolution of cytokine release syndrome. These toxic effects were attributed to other hepatotoxic drugs that she was receiving.
The side-effect profile of multiple infusions was similar to the side-effect profile of a single infusion.
The Safety Switch: How the Suicide Gene Worked
One patient received two infusions of rimiducid. Rimiducid is a dimerizing agent intended to induce caspase 9 and destroy the transferred GD2-CART01 cells. This patient received rimiducid after an altered state of consciousness developed. However, the diagnostic workup revealed a brain hemorrhage as the causative event.
The researchers evaluated the kinetics of GD2-CART01 over time in the patient who received rimiducid. They observed a sharp decrease in the level of circulating GD2-CART01 4 hours after infusion of the dimerizing agent. This decrease was sustained after the second infusion.
After 6 weeks, the researchers detected a new expansion of GD2-CART01. The levels remained elevated at subsequent follow-up and were still detectable 30 months after infusion. The patient continued to have a complete response.
The researchers then evaluated whether GD2-CART01 cells that re-expanded could derive from a subpopulation of cells that had lost sensitivity to rimiducid. GD2-CART01 cells collected at week 8 were exposed in vitro to rimiducid for 96 hours. This exposure produced a remarkable decrease in the number of CAR-positive T cells.
How the CAR T Cells Grew and Survived in the Body
GD2-CART01 cells were detected in vivo by flow cytometry in 26 of 27 patients (96%). Flow cytometry is a laboratory method that counts and sorts individual cells.
The researchers evaluated the correlation between the dose levels tested in phase 1 and the peak GD2-CART01 level in peripheral blood. They used both flow cytometry and droplet digital polymerase-chain-reaction (PCR) assay. Overall, peak expansion was observed in the second week after infusion on flow cytometry. The droplet digital PCR signal anticipated this peak by almost 1 week.
GD2-CART01 persisted for at least 3 months in 75% of the evaluable patients and for more than 2 years in 2 children. The median persistence was 3 months, with a range of 1 to 30 months. GD2-targeted CAR T cells were detectable in peripheral blood in 26 of 27 patients up to 30 months after infusion.
In patients who received a second infusion, assessment of GD2-CART01 persistence after the first dose was terminated at the time of the second dose infusion. This termination occurred regardless of the circulating levels detected. The magnitude of expansion, expressed as an area under the curve, was 329.8 ± 165.7.
Expansion of GD2-CART01 differed between patients who had a response to treatment and those who did not. The mean peak level of GD2-CART01 and the area under the curve in patients with a response were 157.95 ± 97.35 cells per cubic millimeter and 436.0 ± 229.0, respectively. In patients who did not have a response, these values were 70.38 ± 17.90 cells per cubic millimeter and 224.4 ± 126.6, respectively.
After expansion, a predominance of CD8-positive CAR-positive cells was detected in peripheral blood and was maintained at all time points. CD4-positive CAR-positive cells also expanded and were detectable. For both CD4-positive CAR-positive and CD8-positive CAR-positive cells, the phenotype showed a strong prevalence of central and effector memory subpopulations. Central memory CD4-positive cells showed prolonged persistence.
GD2-CART01 cells were detected in bone marrow up to 2 years after infusion. They were detected in the cerebrospinal fluid (the fluid around the brain and spinal cord) in five patients up to 12 weeks after infusion.
The researchers did not observe any difference in bone marrow infiltration by GD2-CART01 between patients with a response and those without a response. The response group included six patients who had a complete response in the bone marrow. The non-response group included six patients who did not have a complete response in the bone marrow and three patients who had relapse in the bone marrow after infusion.
GD2-CART01 cells that persisted long term did not show a relevant expression of exhaustion markers. Exhaustion markers are signs that immune cells have become worn out.
At the end of production, GD2-CART01 cells in the four patients with low in vivo expansion were characterized by a significantly higher proportion of terminally differentiated CD4-positive effector memory cells re-expressing CD45RA. Low in vivo expansion was defined as a peak of expansion less than 5% in peripheral blood. This proportion was higher than in the products of patients with higher in vivo expansion.
Treatment Responses: How Well It Worked
Seventeen children had a response to the treatment, for an overall response of 63%. Nine patients had a complete response, and 8 had a partial response.
Five of 27 patients (19%) had stable disease, and 5 of 27 patients (19%) did not have a response. A total of 17 of 27 patients (63%) had an objective response.
In the entire cohort, the MIBG score decreased from a mean of 13 (range, 0 to 72) to 8.
Survival Results
Among the patients who received the recommended dose, the 3-year overall survival was 60%. The 3-year event-free survival was 36%.
Overall survival means the percentage of patients still alive at that time point. Event-free survival means the percentage of patients who remained alive without a relapse, progression, or other defined event.
Three Patients With an Unusual and Encouraging Course
In three patients (Patients 18, 42, and 44), the researchers observed a peculiar clinical course. All three patients had a reduction in MIBG-positive lesions that was consistent with a partial response after the first infusion.
Two of them (Patients 18 and 42) received additional infusions of GD2-CART01. One then had stable disease, and a deeper partial response developed in the other.
In the subsequent follow-up, the three patients had persistently stable disease in the long term, although no other therapy was administered. All three patients had a negative positron-emission tomographic scan. The researchers noted that the durability of these remissions will be informative.
What These Findings Mean for Patients
The use of GD2-CART01 was feasible and safe in treating high-risk neuroblastoma. Treatment-related toxic effects developed, and the activation of the suicide gene controlled side effects. GD2-CART01 may have a sustained antitumor effect.
For families facing relapsed or refractory high-risk neuroblastoma, these results offer a meaningful signal of hope. This group of children had a very low probability of recovery with subsequent treatments before this trial. The long-term survival in this situation is approximately 5 to 10%.
In this trial, 63% of patients had an objective response. Among patients who received the recommended dose, 60% were alive at 3 years and 36% were alive without an event at 3 years.
The safety switch worked as designed. In the one patient in whom it was activated, the transferred cells were rapidly eliminated. The cells later re-expanded, and the patient continued to have a complete response.
The researchers also found that higher expansion of GD2-CART01 in the blood was associated with response to treatment. Patients with a response had a mean peak level of 157.95 cells per cubic millimeter, compared with 70.38 cells per cubic millimeter in patients without a response.
Importantly, no central neurotoxic effects developed in any patient. This was a key concern that the suicide gene was designed to address.
Limitations of the Study
This was a small, single-center, phase 1–2 trial with 27 patients. Small trials cannot prove that a treatment is better than another treatment. They can show that a treatment is feasible, safe, and worthy of further study.
The trial did not include a comparison group. Without a comparison group, researchers cannot be certain how much of the benefit is due to the treatment itself.
Severe hematologic toxic effects developed in all patients. These effects were induced by lymphodepletion and were sustained after GD2-CART01 infusion. Patients receiving this treatment need close monitoring of their blood counts.
Some patients had hepatic toxic effects, and in some cases these were attributed to other hepatotoxic drugs or to disease infiltration of the liver. This makes it difficult to attribute every liver effect to GD2-CART01 alone.
The researchers noted that the durability of the remissions in the three patients with an unusual course will be informative. Longer follow-up is needed to understand how long responses last.
What Patients and Families Should Know
GD2-CART01 is an investigational therapy. It was tested in a clinical trial, and it is not yet a standard treatment option. Families interested in this approach should ask their treatment team about available clinical trials.
The trial enrolled children and young adults between 1 and 25 years of age. All patients had high-risk neuroblastoma that had relapsed or that had not responded to first-line treatment.
Before receiving GD2-CART01, patients received lymphodepleting chemotherapy with fludarabine and cyclophosphamide. This step is required to make room for the CAR T cells.
Families should expect that side effects can occur. The most common drug-related side effect was cytokine release syndrome, which occurred in 74% of patients. In 95% of those cases, it was mild. Severe blood cell shortages occurred in all patients.
Families should also know that the treatment includes a built-in safety switch. If dangerous side effects occur, doctors can activate this switch to eliminate the transferred cells.
Patients in this trial could receive up to four infusions. Eleven patients received multiple infusions. The side-effect profile of multiple infusions was similar to that of a single infusion.
Questions families may want to ask their care team include:
- Is my child eligible for a GD2-CAR T-cell clinical trial?
- What tests are needed before treatment, such as MIBG scans and bone marrow tests?
- What side effects should we watch for after infusion?
- How long will my child need to be monitored after treatment?
- What options are available if the treatment does not work?
Frequently Asked Questions
What is GD2-CART01 and how does it work?
GD2-CART01 is an investigational CAR T-cell therapy made from a patient's own T cells. It reprograms them to recognize and attack cells carrying the GD2 molecule, which neuroblastoma cells have in high levels. It also includes a built-in safety switch that doctors can activate to eliminate the transferred cells if dangerous side effects occur. It is not yet a standard treatment option.
Who was eligible for this trial?
The trial enrolled children and young adults aged 1 to 25 years with high-risk neuroblastoma that had relapsed or had not responded to first-line treatment. In the phase 2 portion, patients with metastatic disease and MYCN amplification could enroll after first-line treatment, even if they had a complete response. All patients had disease resistant to at least two prior lines of treatment.
What side effects occurred with GD2-CART01?
Cytokine release syndrome occurred in 20 of 27 patients (74%), but it was mild in 19 of those 20 (95%). Severe blood cell shortages developed in all patients due to lymphodepleting chemotherapy. No central neurotoxic effects occurred. Other side effects included liver enzyme elevations, rash, and one brain hemorrhage. Multiple infusions had a similar side-effect profile to a single infusion.
How well did the treatment work in this trial?
Seventeen of 27 patients (63%) had an objective response, including 9 complete responses and 8 partial responses. Among patients who received the recommended dose, 3-year overall survival was 60% and event-free survival was 36%. This was a small, single-center trial without a comparison group, so these results cannot prove the treatment is better than another treatment.
What is the safety switch and did it work?
The safety switch is a built-in suicide gene called inducible caspase 9. If dangerous side effects occur, doctors can give a drug called rimiducid to activate the switch and destroy the transferred cells. In one patient who received it, the cells dropped sharply within 4 hours. The cells later re-expanded, and the patient continued to have a complete response.
What does 3-year overall survival of 60% mean?
Among patients who received the recommended dose, 3-year overall survival of 60% means that 60% of those patients were still alive three years after treatment. Event-free survival of 36% means 36% were alive without relapse, progression, or another defined event. These figures come from a small trial of 27 patients and need confirmation in larger studies.
What should families expect during treatment?
Before receiving GD2-CART01, patients receive lymphodepleting chemotherapy with fludarabine and cyclophosphamide to make room for the CAR T cells. Side effects can occur, most commonly cytokine release syndrome, which was mild in most cases. Severe blood cell shortages occurred in all patients, so close monitoring of blood counts is needed. Patients could receive up to four infusions.
My child has relapsed high-risk neuroblastoma — when should we seek a second opinion about CAR T-cell therapy?
Relapsed or refractory high-risk neuroblastoma has a long-term survival of approximately 5 to 10% with subsequent treatments, so families may want independent review before committing to a plan. GD2-CART01 is investigational and not yet a standard option; it was tested in 27 children, with a 63% objective response and, at the recommended dose, 60% overall survival at 3 years. A second opinion can confirm eligibility, prior GD2 antibody treatment, MIBG score, and whether a clinical trial fits. Diagnostic Detectives Network provides independent expert second opinions.
Source Information
Original article title: GD2-CART01 for Relapsed or Refractory High-Risk Neuroblastoma NEJM 2023
Authors: F. Del Bufalo, B. De Angelis, I. Caruana, G. Del Baldo, M.A. De Ioris, A. Serra, A. Mastronuzzi, M.G. Cefalo, D. Pagliara, M. Amicucci, G. Li Pira, G. Leone, V. Bertaina, M. Sinibaldi, S. Di Cecca, M. Guercio, Z. Abbaszadeh, L. Iaffaldano, M. Gunetti, S. Iacovelli, R. Bugianesi, S. Macchia, M. Algeri, P. Merli, F. Galaverna, R. Abbas, M.C. Garganese, M.F. Villani, G.S. Colafati, F. Bonetti, M. Rabusin, K. Perruccio, V. Folsi, C. Quintarelli, and F. Locatelli, for the Precision Medicine Team–IRCCS Ospedale Pediatrico Bambino Gesù.
Publication: The New England Journal of Medicine, 2023;388:1284-95. DOI: 10.1056/NEJMoa2210859.
Trial registration: ClinicalTrials.gov number NCT03373097.
Funding: Funded by the Italian Medicines Agency and others.
This patient-friendly article is based on peer-reviewed research.
