Background
T-lineage acute lymphoblastic leukemia (T-ALL) is an aggressive malignancy derived from early T cell progenitors and constitutes 10% of childhood ALL and 20% of adult ALL cases [
1‐
3]. Despite recent advances in treating the disease, 30% of T-ALL cases are resistant to or relapse after front-line chemotherapy regimens. Allogeneic stem cell transplantation (SCT) has been recommended as a salvage option for these patients, but only those who can be re-induced to remission are eligible [
4]. Overall, the prognosis for patients with relapsed or refractory disease is very dismal due to the lack of new treatment options [
5‐
11].
Following the development of a line of chimeric antigen receptor (CAR) T-cell therapies for B cell leukemias, there is growing interest in developing novel cell-based therapies for T-ALL [
12]. However, developing therapeutic CAR T-cell treatments for T-ALL is particularly challenging because most validated CAR targets are also expressed on normal T cells. This results in several obstacles to CAR T-cell therapy in T-ALL, including the risk of tumor contamination, fratricide of CAR T cells, and the depletion of healthy T cells, which may make patients more susceptible to opportunistic infection [
13]. Despite these hurdles, recent preclinical findings have prompted the initiation of early clinical trials of CAR T-cell therapies for patients with r/r T-ALL [
14,
15].
We previously reported early results from a phase I first-in-human study showing that CD7-directed CAR T cells manufactured from T cells collected from allogeneic donors with retention of CD7 molecules in the endoplasmic reticulum could partially overcome these barriers to treat r/r T-ALL [
16]. The primary analysis showed that grade 3–4 cytokine release syndrome (CRS) and grade 1 or 2 graft-versus-host disease (GVHD) were short-term adverse events (AEs) that occurred < 30 days after treatment in 10% and 60% of patients, respectively. At 1 month after treatment, 85% of patients had a minimal-residual-disease-negative complete response [
16]. Despite these favorable efficacy and safety findings, the short follow-up period of a median of 6.3 months did not allow adequate assessment of remission durability or of long-term AEs, both of which are critical to ascertain the risk–benefit profile.
Here, we report a protocol prespecified long-term analysis of safety and efficacy outcomes in this cohort at a median follow-up time of 27 months. Long-term pharmacokinetics and change of endogenous lymphocyte subpopulations will also be presented.
Discussion
This study provides a 2-year follow-up in 20 participants with r/r T-ALL after therapy with CD7-directed CAR T cells originated from donors. We showed durable remissions lasted for more than 24 months in a proportion of treated patients, whereas relapse emerges as a main cause of treatment failure. This research also raised infection and GVHD as major long-term adverse events in patients without SCT consolidation.
The safety analysis in this follow-up study provides further insights into long-term risks of the therapy. Our previous report demonstrated manageable short-term adverse events, including cytopenias, CRS, neurotoxicity and acute GVHD. Here we showed that late-onset GVHD was the most common long-term AEs with 58% incidence among the 12 patients without SCT consolidation. The allogenic origin of T cells was suspected to be a contributing factor in late-onset GVHD, however these complications were mostly mild and all manageable. Interestingly, our results did not show a significant association between human leukocyte antigen (HLA) matching degree with incidence or persistence of GVHD, but due to the small sample this issue needs to be further investigated in future studies. Only one case of severe GVHD was reported, and no severe GVHD occurred beyond 12 months post infusion, suggesting that this AE was generally manageable.
Grade three or worse infections occurred in 42% (5/12) of participants who did not receive SCT consolidation. These severe infections mostly occurred around 6 months to 1 year post infusion, and were suspected to be caused by mixed effects of normal T-cell depletion, cytopenias, and immunosuppressive agents that were used to control GVHD. During the follow-up period, 58% (7/12) of patients who did not underwent SCT consolidation had their non-CAR T cells restored to normal level. The incidence of severe infections seemed to be negatively correlated with T-cell recovery, although not statistically significant. Long-term monitoring of T cells in some patients showed increased central memory subpopulation and good responsiveness to viral antigen stimulation, suggesting that they had some protective effects. Notably, a lower incidence of severe infections (1/7, 14.3%) were observed in patients with SCT consolidation after CAR T-cell infusion, suggesting that early bridging to SCT may reduce the risk of life-threatening infection. For patients who were planning to receive new donor-derived CD7 CAR T cells, we would recommend them to discard this CAR T-cell therapy if subsequent SCT was not feasible. Patients treated with prior-SCT donor-derived CD7 CAR T cells mostly had available donors and were also strongly recommended for SCT consolidation, but most of them declined it for personal reasons. The future effort will be made to arrange most patients to accept SCT after CAR T-cell infusion. For patients who are not eligible for SCT, using a molecular switch to terminate CAR T cells, or early infusion with purified CD34
+ stem cells to promote T-cell recovery, may be the alternative strategies to reduce the risk of severe infections [
20]. Nonetheless, the relationship between T-cell recovery and risk of severe infection, as well as the ways to overcome this challenge, remains to be further investigated.
The treatment produced a median PFS of 11.0 months and a median OS of 18.3 months, and 31.5% (
n = 6) of the responders achieved responses lasting more than 2 years, including three who did not receive further consolidation. A previous study reported a median OS of 8 months among patients with r/r T-ALL treated with salvage nelarabine treatment [
21]. Of the six (33.3%) patients who relapsed, four were CD7-negative relapse and two were CD7-positive relapse. CD19-negative relapse frequently occurred during CD19 CAR T-cell therapy for B-cell malignancies, with various mechanisms including mutation of CD19 gene and evolution from pre-existing CD19-negative subclones [
22,
23]. Frameshift or missense mutations were detected in the four CD7-negative relapse patients in our study, suggesting that mutation may be a main cause of CD7 loss in tumor cells. However, due to the lack of sufficient samples, it remains to be determined in future research whether CD7-negative relapse was derived from CD7-negative leukemic (or preleukemic) clones that exist before therapy, or caused by a new mutation during CAR T-cell treatment. We also observed CD7-positive relapse following the loss of CAR T cells (at flow cytometric level) in a patient without SCT consolidation, suggesting that insufficient persistence may also be a cause of relapse. However, the incidence of this antigen-positive relapse is relatively low, consistent with the overall good CAR T-cell persistence that are possibly related to the fine CAR vector design and donor complete chimerism status.
The CD7-negative T cells after CAR T-cell therapy could be derived from endogenous CD7-negative T cells (naturally developed) preexisted in patients, or from the infused T-cell product (naturally developed or caused by endoplasmic reticulum retention of CD7 protein). The eight patients who received new donor CAR T cells obtained complete or mixed chimerism status at 1 month after CAR T-cell infusion, suggesting that their CD7-negative T cells were mostly or partially derived from the T-cell product. The 12 patients who received prior-SCT donor-derived CAR T cells had already achieved complete chimerism after SCT and contain donor stem-cell differentiated CD7-negative T cells, therefore it could not be determined whether their CD7-negative T cells were derived from endogenous T cells or infused T-cell product. The clear dissection of the origin and function of these CD7-negative T cells warrants further analysis.
Differences in study designs and patient characteristics make it difficult to directly compare the results of donor CD7 CAR T-cell therapy with the recent reports of other ongoing CD7-targeted cellular therapies. Genome-edited universal CD7 CAR T cells, autologous nanobody-based CD7 CAR T cells and naturally selected CD7 CAR T cells also showed early efficacy in patients with T-ALL, while CD7-negative relapses were also observed in these studies [
16,
24‐
26]. Fewer severe infection cases were reported in these studies. However, these therapies had a less than 1-year median follow-up or were mostly bridged to SCT early after CAR T-cell infusion, making it difficult to evaluate the risk of infection with the long-term presence of CAR T cells. In contrast, our patients were followed up for more than 2 years, and more than half of them did not receive SCT post CAR T-cell infusion, allowing us to assess the long-term risk of infection. Donor-derived CD7 CAR T cells may be particularly useful, when autologous CD7 CAR T-cell therapy is unavailable due to low quantity or quality of patient’s T cells, or risk of tumor contamination. Indeed, the preparation of autologous T cell for patients with T-ALL in some patients may be very challenging, since the chemotherapy regimen were usually very intense and designed against T-lineage cells [
1]. A parallel phase I trial is undertaken in our center to test feasibility of autologous CD7 CAR T cells against T-ALL with the same lentiviral vector, and this will provide further information to compare the application between donor and autologous CD7 CAR T cells [
27].
This study has several limitations. Firstly, it was a phase I trial designed to initially explore feasibility and safety with a small sample size. Therefore, a phase II study with a larger sample size is needed to confirm safety and efficacy and to define prognostic factors. In addition, it is the first experience with donor CD7 CAR T cells for T-ALL, with no prior clinical experience, and regimens for managing adverse events need to be further optimized. Also, long-term functionality of non-CAR T cells and the mechanisms of CD7 antigen loss that leads to relapse warrants further investigation.
The main strength of this study is that it presents a first long-term follow-up in patients with T-ALL after CAR T-cell treatment. The durability of responses and new signals of long-term adverse events showed here, as well as previous report of a high early response rate, support that donor-derived CD7 CAR T cells may be a feasible salvage treatment for children or adults with r/r T-ALL. Severe infection appears to the notable side effect associated with this therapy. Early bridging to SCT consolidation has the potential to reduce this risk of infection, and monitoring of immune function and careful prevention and treatment of infection is critical for those patients for whom a subsequent SCT is not feasible. To know more about the benefit and risk profile, a multicenter phase II clinical trial is ongoing, which may provide additional clues for further optimization of this therapy.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.