CD7 CAR-T therapy: current developments, improvements, and dilemmas

Abstract

Chimeric antigen receptor (CAR) T-cell therapy is an epoch-making immunotherapy for the treatment of relapsed or refractory (r/r) blood tumors, as demonstrated by its successful implementation in r/r B cell-derived malignancies. However, replicating this success in T-cell leukemia or lymphoma remains challenging. Among the various potential target antigens, CD7 has garnered attention as a promising candidate. CD7 CAR-T therapy is one of the most extensively studied approaches for treating r/r T-cell acute lymphoblastic leukemia/lymphoblastic lymphoma (T-ALL/LBL) and r/r acute myeloid leukemia (AML). Based on the source of T cells, CAR-T products can be categorized as autologous and allogeneic, both of which are being tested in clinical trials, each offering specific advantages. Allogeneic CD7 CAR-T cells outperform autologous cells in terms of reducing manufacturing costs, ensuring consistent quality, and improving affordability and availability. Despite these advantages, challenges like graft-versus-host disease (GVHD), host-versus-graft reaction (HVGR), and fratricide pose significant barriers to the clinical application of allogeneic CD7 CAR-T cells. However, innovative gene-editing techniques, such as CRISPR/Cas9 and base editing, and more promising cell sources, such as natural killer T (NKT) cells and induced pluripotent stem cells (iPSCs), are emerging as potential solutions. In this review, we discuss the different categories of CD7 CAR-T products, their application in clinical settings, and directions for refinement.

1. INTRODUCTION

T-cell acute lymphoblastic leukemia (T-ALL) is a highly aggressive hematological malignancy that arises from the oncogenic transformation of T-cell precursors.^1,2 As a subset of ALL, T-ALL represents 10% to 15% of pediatric ALL cases and 20% of adult ALL cases.³ Currently, chemotherapeutic agents and hematopoietic stem cell transplantation (HSCT) remain the dominant therapeutic strategies for T-ALL. Although the 5-year overall survival (OS) rate exceeds 80% in pediatric patients with T-ALL, the outcome is less favorable in adults, with a 5-year OS rate <50%.^3,4 The disease becomes particularly intractable when it relapses or is refractory to initial chemotherapy. Most patients are ineligible for salvage HSCT because of failure to reinduce remission.⁵ Consequently, the outlook for individuals with relapsed or refractory (r/r) T-ALL continues to be bleak due to the absence of effective new treatments.^3,4,6 With the advancement in cellular immunotherapy, chimeric antigen receptor (CAR) T-cell therapy has shown promise in treating patients with r/r B-ALL/lymphoblastic lymphoma (LBL).^7,8 To date, 7 autologous CAR-T therapies have been approved by the United States Food and Drug Administration (US FDA). Inspired by the successful application of CAR-T cells in B-cell malignancies, efforts are underway to extend this approach to treat T-ALL/LBL. Numerous studies have explored potential T-lineage-specific antigens that can be targeted, such as CD5, CD7, and CD3.⁹ Among these, CD7 is the most commonly investigated target. CD7 is a transmembrane glycoprotein expressed in more than 95% of T-ALL and T-LBL cases, as well as in a subset of peripheral T-cell lymphomas.¹⁰ In healthy individuals, CD7 is expressed in 90% to 96% of normal T cells and 90% to 98% of natural killer (NK) cells.^11,12 Intensive expression of CD7 in tumor cells underpins targeted killing by T cells engineered to express a CD7-targeting CAR. However, the presence of CD7 in both healthy blood and leukemic cells triggers a range of unintended consequences of CD7 CAR-T therapies. This overlap can negatively affect the production and overall quality of CAR-T cells, potentially resulting in severe or fatal outcomes.¹³

In this review, we examine the main types of CD7 CAR-T therapies, their principal advantages and limitations (Table 1), clinical application scenarios (Table 2), and future directions, as implied in ongoing CD7 CAR-T clinical trials.

Table 1.

Comparison of autologous and allogeneic CAR-T cells: advantages and disadvantages.

	Autologous CAR-T cells	Allogeneic CAR-T cells
Derivation of the starting material	Patients	Healthy individuals
Advantages	No risk of GVHD and HVGR; Intermediate to long persistence (months to years) in vivo	Possibility of industrialized production and consequently reduced costs and improved affordability; Optimal strictly selected substrate T cells with prominent efficacy; No concerns over product contamination; Immediate availability for rapidly progressing cases
Disadvantages	High financial and time manufacturing costs; Potential of leading to postponement of reinfusion; Product contamination with leukemic cells; Unstable donor-dependent quality and efficacy of the substrate T cells and final product	Possibly causing GVHD and HVGR; Extra genetic-engineering steps to prevent Short to intermediate persistence (weeks to months) after reinfusion

CAR = chimeric antigen receptor, GVHD = graft-versus-host disease, HVGR = host-versus-graft disease.

Table 2.

Published clinical trials of CD7 CAR-T cells.

Clinical trials of CAR-T cells	Strategies	Patients	Efficacy and toxicities
Autologous nanobody–derived fratricide-resistant CD7-CAR-T: Phase I	PEBL: downregulate surface CD7	8 patients: 1 with T-ALL, 1 with mixed-phenotype acute leukemia, 6 with T-ALL/LBL	Efficacy: Of 8 patients, 7 achieved CR (75.0%), of 6 T-ALL/LBL patients with BM involvement, 6 achieved BM MRD negative CR; CRS: in 8 (100%, grade 1–2 in 7 [87.5%]); ICANS or GVHD: none; Cytopenias: in 8 (100%; grade 4 lymphopenia in 8 [100%])14
CD7 CAR-T: Phase I	PEBL: downregulate surface CD7	17 patients with T-ALL	Efficacy: Of 17 patients, 16 achieved MRD-negative CR in PB and BM (94.1%); CRS: in 13 (76.5%; all grade 1–2); ICANS: in 2 (11.8%; all grade 1); Cytopenias: in 17 (100%; grade 3–4 ICAHT in 13(76.5%))15
NS7 CAR-T: Phase I	Natural selection: downregulate available surface CD7	20 patients, 14 with T-ALL, 6 with T-LBL	Efficacy: Of 20 patients, 19 achieved BM MRD-negative CR/CRi (95%); CRS: in 19 (95%; grade 1–2 in 18 [90%]); ICANS: in 2 (10%; all grade 1); Cytopenias: in 20(100%; grade 4 lymphopenia and neutropenia in 20 [100%])16
NS7 CAR-T: Phase I/II	Natural selection: downregulate available surface CD7	60 patients: 35 with T-ALL, 25 with T-LBL	Efficacy: Of 54 patients with BM blast infiltration, 51 achieved MRD-negative CR (94.4%); ORR: 96.3%; 1-y PFS: 58.6%; 2-y PFS: 53.7%; 1-y OS: 72.7%; 2-y OS: 63.5%; CRS: in 55 (97%; grade ≥3 in 11 [11.7%]); ICANS: in 3 (5%; grade 1 in 2 [3.3%], grade 4 in 1 [1.7%]); Grade ≥3 Cytopenias: lasting ≥30 d in 37 (61.7%); lasting ≥60 d or present at HSCT in 16 (26.7%) Infections: in 22 (36.7%; grade 3 in 10 [16.7%])17
GC027: Phase I	CRISPR–Cas9: knock out CD7 and TRAC	12 adults: 11 with T-ALL, 1 with T-LBL	Efficacy: Of 12 patients, 11 achieved CR/CRi (91.7%); CRS: in 10 (83.3%; grade ≥3 in 8 [66.7%]); ICANS: none; GVHD or cytopenias: NR18
RD13-01: Phase I	CRISPR–Cas9: knock out CD7, TRAC RFX5 Incorporate iNK and γC	12 adults: 7 with T-ALL, 2 with T-LBL, 1 with NKTCL, 1 with PTCL, NOS, 1 with AML	Efficacy: Of 11 evaluable patients, 7 achieved CR/CRi (63.6%); CRS: in 10 (83.3%; all grade 1–2); ICANS or GVHD: none; Cytopenias: NR19
BE-CAR7: Phase I	Base editing: knock out TRAC, CD52, CD7	3 pediatric patients: 3 with R/R T-ALL	Efficacy: Of 3 patients, 3 achieved CR/CRi (100%); CRS: in 3 (100%; grade 4 in 1 [33.3%]); ICANS: in 1 (33.3%; low grade); GVHD: in 1 (33.3%; mild); Cytopenias: in 3 (100%; all grade ≥3)20
CD7 CAR-T: Phase I	IntraBlock technology: downregulate surface CD7	20 patients with T-ALL	Efficacy: Of 20 patients, 18 achieved CR (90%); (<30 d) CRS: in 20 (100%; grade 1–2 in 18 [90.0%]); ICANS: in 3 (15.0%; all grade 1–2); GVHD: in 12 (60%; all grade 1–2); Cytopenias: in 20 (100%; all grade 3–4); Viral reactivation: in 4 (20.0%; all grade 1–2)21
CD7 CAR-T: Phase I	IntraBlock technology: downregulate surface CD7	20 patients with T-ALL, 12 with CD7 CAR-T only	Efficacy: Of 12 patients with CD7 CAR-T only: 2-y PFS: 31.8%; 2-y OS: 35%; median PFS: 11.0 mo; median OS: 18.3 mo; (>30 d) Of the 12 patients without SCT consolidation: CRS or ICANS: none; GVHD in 7 (58.3%; grade 1–2 in 6 [50%]); Cytopenias: Late-onset cytopenias in 3 (25%); Infection: in 6 (50%; ≥grade 3 in 5 [41.7%])22
NS7CAR-T: Phase I	Natural selection: downregulate available surface CD7	10 patients with r/r AML	Efficacy: Of 10 patients, 7 achieved CRi (70%; MRD-negative CRi in 6 [60%]); CRS: in 10 (100%; grade 1–2 in 8 [80%]); ICANS: none; Cytopenia: grade 4 lymphopenia and neutropenia in 10 (100%), grade 4 thrombocytopenia in 9 (90%)23

BM = bone marrow, CAR = chimeric antigen receptor, Cas9 = CRISPR-associated protein 9, CR = complete remission, CRi = complete remission with incomplete hematological recovery, CRISPR = clustered regularly interspaced short palindromic repeats, CRS = cytokine release syndrome, GVHD = graft-versus-host disease, ICANS = immune effector cell-associated neurotoxicity syndrome, MRD = minimal residual disease, NOS = not otherwise specified, NR = not reported, PEBL = protein expression blocker, r/r = relapsed/refractory, T-ALL = T cell acute lymphoblastic leukemia, T-LBL = T cell lymphoblastic lymphoma.

2. CATEGORIES OF CD7 CAR-T CELLS AND THEIR APPLICATION SCENARIOS

Based on the source of the substrate T cells, CAR-T therapy can be categorized into 2 types: autologous and allogeneic. Allogenic CAR-T products can be further divided into HLA-matched or haploidentical, and HLA-unmatched donor-derived CAR-T products.²⁴ Clinically, whether autologous or allogeneic CAR-T therapy is adopted depends on the patient’s condition and the clinicians’ evaluation of which option offers a better risk-benefit profile. In addition to its expression in T-ALL/LBL and ETP-ALL, the most aggressive subtype of T-ALL,^10,25 CD7 is found in up to 24% of acute myeloid leukemia (AML) cases and is thought to be a marker of leukemic stem cells.^26,27 Consequently, CD7 CAR-T cell therapy has emerged as a promising immunotherapy for CD7-positive, r/r T-ALL/LBL, and r/r AML patients, who are otherwise devoid of efficacious treatment and have dismal prognosis.

2.1. Autologous CD7 CAR-T cell products

To date, all 7 FDA-approved CAR-T products are autologous, demonstrating the feasibility of autologous CAR-T therapy in clinical settings. Notably, autologous CAR-T cells exclude the risk of GVHD or HVGR after reinfusion, which are major hurdles for allogeneic CAR-T cells.²⁸ These issues will be discussed in more detail later. Multiple studies have verified the strong anti-tumor effects of autologous CD7 CAR-T therapies.

Zhang et al¹⁴ launched a Phase I clinical trial to assess the efficacy and safety of autologous CD7 CAR-T cells in patients with r/r T-ALL/LBL (NCT04004637). The results showed durable anti-tumor responses with slight toxicity. By connecting a tandem CD7 nanobody to an endoplasmic reticulum (ER)/golgi retention signal peptide, the CD7 CAR-T cells were made functionally CD7-negative and fratricide-resistant. Among the 8 enrolled patients with r/r T-ALL/LBL, 7 achieved complete remission (CR) within 3 months after infusion, attaining an 87.5% CR rate, with 1 patient maintaining CR for over 12 months. The majority of participants (87.5%) experienced only grade 1 or 2 cytokine release syndrome (CRS), with no T-cell hypoplasia or neurologic toxicities, indicating a tolerable safety profile.¹⁴ Lu et al¹⁶ recruited 20 r/r/ T-ALL/LBL patients in their first-in-human Phase I trial of naturally selected CD7 CAR-T (NSCAR7T) cells, without gene-editing (NCT04572308). Eighteen patients received autologous CAR-T cells, and the remaining 2 received CAR-T products from their respective donors. The results revealed that 17 patients (85%) who received autologous NSCAR7T cells achieved bone marrow (BM) minimal residual disease (MRD)-negative CR or CR with incomplete hematological recovery (CRi) before day 28. Five of the 9 patients achieved extramedullary CR. Toxicities were generally manageable, with 19 patients experiencing no or slight (≤grade 2) CRS, while 2 patients experienced grade 2 immune effector cell-associated neurotoxicity syndrome (ICANS).¹⁶ Given the comparatively small sample size and short follow-up time, Lu et al¹⁶ initiated the Phase II study (NCT04916860) with extended observation and more participants, based on the Phase I findings. Among the 60 participants, 58 received autologous CD7 CAR-T cells. In 54 patients with BM involvement, 94.4% (51/54) achieved MRD-negative CR in BM/PB. With a median follow-up of 368.5 days, the 2-year OS rate was 63.5%, while the progression-free survival (PFS) rate was 53.7%. Major long-term adverse events (AEs) included CRS, ICANS, and immune effector cell-associated hematotoxicity (ICAHT). CRS was observed in 91.7% (55/60) of patients, with 80% experiencing grade 1/2 CRS and 11.7% experiencing grade 3/4 CRS. Only 5% (3/60) of patients developed ICANS, and 61.7% (37/60) experienced cytopenia 1 month post-infusion, though more than half recovered within 2 months of CAR-T transfusion.¹⁷ Overall, both Phase I and II trials revealed that NSCAR7T therapy was effective in controlling leukemia progression in heavily pretreated patients.

2.2. Allogeneic CD7 CAR-T cells derived from HLA-unmatched donors

Developing universal off-the-shelf CD7 CAR-T cells has long been the mainstream focus of immunotherapy due to their anticipated accessibility, availability, and remarkably lowered manufacturing costs (Table 1).

Cooper et al²⁹ successfully constructed fratricide-resistant off-the-shelf CD7 CAR-T cells by knocking out CD7 and TRAC using the CRISPR Cas9 system. They validated the potent tumor-killing effects on T-ALL cell lines and primary T-ALL cells both in vivo and in vitro. Their preclinical data strongly supported further clinical research of universal fratricide-resistant CD7 CAR-T cells.²⁹ Li et al¹⁸ developed an innovative universal allogeneic CAR-T cell therapy known as GC027, which utilized CRISPR/Cas9 to disrupt the genes encoding the T-cell receptor α (TRAC) and CD7. Additionally, an enhancer derived from IL7Ra was incorporated into the therapy’s design. They enrolled 12 patients with r/r T-ALL/LBL to investigate the efficacy and safety of GC027 in a Phase I clinical trial (ISRCTN19144142), which eventually achieved a CR rate as high as 91.7%.¹⁸ In a Phase I clinical trial (NCT04538599), Hu et al¹⁹ included 12 participants: 7 with T-ALL, 4 with T-LBL, and 1 with AML, administering a genetically modified CD7 CAR-T treatment. The final product, termed RD1301, was constructed by disrupting the encoding genes of CD7, TCR, and HLA-II via CRISPR/Cas9, while also integrating an NK cell inhibitor (NKi) and the common cytokine receptor gamma chain (γc). Of the 11 patients who received CAR-T cell infusion and were eligible for efficacy evaluation, 7 achieved CR or CRi, attaining a complete response rate of 63.6%. After a median follow-up period of 10.5 months, 4 patients remained in CR. None of the participants experienced dose-limited toxicities (DLTs), GVHD, ICANS, or severe CRS (grade ≥3), but all developed grade 4 neutropenia. Notably, several patients experienced viral reactivation, with 1 patient succumbing to Epstein-Barr virus (EBV)-related diffuse large B-cell lymphoma (DLBCL). Overall, RD1301 demonstrated a promising efficacy and safety profile.¹⁹

Base editing is another emerging gene-editing tool that efficiently creates single-base pair conversions at specific genomic target sites.^30,31 Cytosine base editors (CBEs) operate by catalyzing the conversion of C-G to T-A point mutations through cytidine deamination at defined genomic loci, without creating double-strand break (DSBs).³² With the assistance of CBEs, Diorio et al³³ manufactured allogeneic 7CAR8, a quadruple-edited CAR-T product targeting CD7, in which genes encoding CD7, CD52, PD1, and TRAC were silenced. They demonstrated that 7CAR8 was effective in eradicating malignant T cells in multiple models.³³ In a Phase I clinical trial (ISRCTN15323014), Chiesa et al²⁰ employed CRISPR-guided cytidine deamination to disrupt the expression of TRBC, CD7, and CD52, generating BE-CAR7, a universal, off-the-shelf CD7 CAR-T product. In their early report, 3 pediatric patients with r/r T-ALL were administered BE-CAR7, with 2 achieving molecular remission and subsequently becoming eligible for HSCT. Short-term AEs were similar to those observed in other CD7+ CAR-T clinical trials. The study is ongoing, with plans to recruit more patients.²⁰

2.3. Allogeneic CD7 CAR-T cells derived from HLA-matched or haploidentical donors

For patients with a history of HSCT or those with available matched donors, the strategy of using HLA-matched or haploidentical donor-derived CAR-T cells offers an alternative therapeutic approach. Pan et al²¹ conducted the first-in-human Phase I clinical trial of CD7 CAR-T cells in patients with r/r T-ALL (NCT04689659). The study included 20 patients, of whom 14 received haploidentical donor-derived CAR-T cells, and 6 received matched sibling donor (MSD) or matched unrelated donor (MUD) CAR-T cells. The trial achieved a CR rate of 90% (18/20), with CD7 CAR-T cells demonstrating robust expansion and effective anti-leukemia activity in vivo. Seven patients proceeded to HSCT in remission. Of the 18 patients who achieved CR, 15 (83%) maintained remission at a median follow-up of 6.3 months. AEs were overall mild, including grade 1 to 2 CRS (90%), grade 3 to 4 CRS (10%), grade 3 to 4 cytopenia (100%), grade 1 to 2 neurotoxicity (15%), and grade 1 to 2 GVHD (60%).²¹ Encouraged by the favorable efficacy and safety results from the Phase I trial, the same group conducted a 2-year follow-up Phase II trial (ChiCTR2000034762). In this trial, the 2-year PFS rate for the 19 responsive patients was 36.8%, whereas the overall OS rate was 42.3%. CD7 CAR-T cells provided prolonged anti-leukemic benefits, demonstrated by an 11-month median PFS and an 18.3-month median OS. All 12 patients who received donor-derived CAR-T cells experienced resolution of cytopenia with appropriate medications. Among the 12 patients not bridged to HSCT, 7 developed late-onset GVHD, most of which were mild (6/7), and these cases were managed successfully with treatment. However, 5 severe infections raised concerns as a late-onset AE.²²

2.4. CD7 CAR-T cells in AML

CD7 is a pan-T lineage cell surface marker expressed extensively in both normal and malignant T and NK cells.^10–12 While CD7 expression is observed in approximately 20% to 35% of AML cases, it is absent in normal myeloid and erythroid cells. In CD7-positive AML patients, CD7 expression is often associated with poor prognosis.²⁷ Most AML CAR-T products under evaluation target antigens specific to myeloid lineage cells, such as CD33, CD123, and CLL-1.^34–38 Although promising anti-leukemic efficacy has been observed, shared antigenicity between AML cells and normal myeloid cells routinely breeds long-term or even permanent suppression of myeloid hematopoiesis.^39–41 Consequently, an increasing number of researchers are setting their sights on CD7 CAR-T cells as a potential treatment for patients with CD7+ AML. In preclinical trials by Lu et al⁴² and Gomes-Silva et al,⁴³ CD7 CAR-T cells exhibited potent and selective cytotoxic effects on AML cell lines, primary AML blasts in vivo, and in CD7+ AML xenograft models, while sparing hematopoietic stem cells and other myeloid cells. However, the application of CD7 CAR-T in r/r AML patients is not as extensive compared to its use in r/r T-ALL/LBL. Limited clinical data suggest that CD7 CAR-T cells may induce CR in AML patients without causing severe toxicities.^19,23,44 After Lu et al⁴² validated the potent leukemia-eradicating efficacy of naturally selected CD7 CAR-T cells in a preclinical trial, Lu et al²³ combined the strategy of natural selection and anti-CD7 nanobody to generate dual variable heavy (VH)-chain domain of a heavy-chain antibody (dVHH) NS7CAR-T cells. This product was administered to 10 patients with r/r AML. Interestingly, dVHH NS7CAR-T cells induced CRi in 7 (70%) patients within 4 weeks after infusion, 6 of whom were MRD-negative. Two patients developed grade 3 CRS and none experienced ICANS, indicating a tolerable safety profile. Although most patients succumbed to death or relapsed, this Phase I clinical trial justified the integration of consolidative HSCT after attaining CR/CRi to some extent.^23,42 In a clinical case reported by Cao et al,⁴⁴ a patient resistant to multiple lines of chemotherapy achieved morphological CR after treatment with CD7 CAR-T cells, providing an opportunity for HSCT. Considering the advantageous CD7 expression profile and the selective leukemia-eliminating efficacy established in preclinical trials, CD7 CAR-T cells offer a promising treatment option for r/r AML patients and merit further exploration.

2.5. Summary for the completed CD7 CAR-T clinical trials

Most CD7 CAR-T products have achieved excellent CR rates, exceeding 80% or even better (Table 2). Although short-term efficacy is promising, long-term survival outcomes are remarkably variable, as exemplified by CR durations ranging from 2 months to over 2 years. This variation can be partially explained by distinct patient characteristics, such as the specific subtype of T-ALL, comorbidities, and previous treatments. Additionally, factors such as the CAR-T cell dosage and subsequent maintenance therapy strategies may partially account for the disparity in CR duration. However, no large-scale randomized clinical trials have yet explored the practical significance of these factors. Almost all patients developed CRS and cytopenia after CAR-T cell infusion. While CRS was generally grade 1 to 2 and manageable, cytopenia was universally grade 3 to 4 and intractable. Severe and prolonged T-cell aplasia substantially increased the risk of opportunistic infections, including EBV or cytomegalovirus (CMV) reactivation, which adversely affected survival and prognosis.^14–23 As the follow-up period extended, a considerable proportion of patients experienced relapsed CD7-negative or positive T-ALL. CD7 negativity in relapsed T-ALL confers natural resistance to CD7 CAR-T cell therapy, making the disease more refractory. Relapsed CD7-negative T-ALL may originate from a subset of CD7-negative leukemic cells that existed prior to CAR-T cell infusion and exhibited a survival predominance after CAR-T cell treatment. Another possibility is that the expansion of the newly emergent CD7-negative leukemic cells, boosted by gene mutations, promoter demethylation, alternative splicing, or transcriptional silencing, may directly lead to relapse.^45,46 Additionally, Hamieh et al⁴⁷ demonstrated that trogocytosis, a process where CAR-T cells extract the target antigen from the surface of leukemic cells and thus obtain increased positivity for the antigen, facilitated antigen-low tumor relapse. When the density of the target antigen declined below the threshold needed for activating CAR-T cells and provoking targeted killing, the risk of relapse rised.⁴⁷ Dual antigen-targeting CAR-T therapy and bridging to HSCT are anticipated to conquer or avoid CD7-negative relapse and are under investigation.

Both autologous and allogeneic CAR-T cells demonstrated a tolerable efficacy and safety profile. In clinical practice, the choice between autologous and allogeneic CAR-T therapies depends on factors such as the patient’s characteristics and the specific product being used (Table 1). Allogeneic CAR-T cells offer key advantages, including ready accessibility and stable quality, making them preferable for immunocompromised patients or those with aggressive disease progression. With immunotherapy now playing a leading role in treatment protocols, a growing number of patients, particularly those with intact immune function and no major comorbidities, are opting for CAR-T therapy over traditional chemotherapy as their consolidation treatment of choice. For individuals deemed unlikely to experience rapid disease progression while awaiting CAR-T cell production, autologous CAR-T treatments present an advantageous option, often outperforming allogeneic therapies by minimizing the risk of immune rejection complications.

3. ISSUES REMAINING TO BE OPTIMIZED

CAR-T cell products have demonstrated the ability to eradicate leukemia in preclinical and clinical trials. However, there is still significant room for improvement. The optimal CAR structure and the most efficient CAR transduction approaches remain to be explored. Additionally, fratricide remains a major challenge, negatively contributing to both productivity and anti-leukemic activity of CAR-T products. In clinical trials, GVHD, HVGR, and T-cell aplasia are common AEs after CAR-T cell reinfusion, which can be confirmed using CAR-T cell modification technologies (Fig. 1).

Figure 1.

Commonly used techniques to avert fratricide, GVHD and HVGR. (A) CRISPR/Cas9: knock out CD7 to prevent fratricide, TRAC or TRBC1 or TRBC2 to avoid GVHD, B2M, or CIITA to circumvent HVGR. (B) Base editing: edit out CD7 to prevent fratricide, TRAC or TRBC1 or TRBC2 to avoid GVHD, B2M, or CIITA to circumvent HVGR. (C) PEBL: anchor CD7 in ER and/or Golgi apparatus to block CD7 trafficking to the T cell membrane. (D) Ibrutinib and dasatinib: inhibit the key kinases in CAR signaling pathway to block CAR cell activation. (E) CAR-NK cells: eliminate tumor cells with autocrine IL-15 to enhance persistence without causing GVHD. (F) CAR-γδT cells: eradicate tumor cells without causing GVHD. (G) iPSCs: reprogrammed from other cell types to differentiate into a wide range of cells for CAR cell production. (H) UCB cells: provide available HSPCs to manufacture CAR-cell products with advantageous phenotypes. B2M = β2 microglobulin, CAR = chimeric antigen receptor, CIITA = class II transactivator, ER = endoplasmic reticulum, GVHD = graft-versus-host disease, HNH = histidine-asparagine-histidine, HSPC HVGR = host versus graft reaction, IL = interleukin, iPSC = induced pluripotent stem cell, MHC = major histocompatibility complex, NK = natural killer, PAM = protospacer adjacent motif, PEBL = protein expression blocker, TRAC = T-cell receptor α, TRBC UCB = umbilical cord blood, βTCR = β T cell receptor.

3.1. Fratricide

As previously outlined, CD7 is co-expressed by both normal and tumor T cells.^10,11 Therefore, the surface expression of CD7 on CAR-T cells can also direct CAR-T cells against each other, provoking mutual recognition and self-attack. Trogocytosis is presumed to be a newly discovered mechanism of fratricide.⁴⁷ Fratricide poses a significant barrier to CAR-T cell proliferation, effectiveness, and longevity.⁴⁸ To address this, several methods have been employed to minimize its impact during the manufacturing of CD7 CAR-T cells.

Genetic deletion of CD7 prevents fratricide among CAR-T cells by making them CD7-negative. Multiple trials have utilized the CRISPR/Cas9 RNP complex to efficiently disrupt CD7 expression and construct fratricide-resistant CD7 CAR-T cells.^18,19,29,49 However, CRISPR/Cas9-based editing relies on the formation of DSBs, raising concerns over genetic aberrations, especially under the circumstances of multiplexed editing.^50–53 Nevertheless, multiplex gene editing is essential for the safe use of allogeneic CAR-T cells, driving research into base-editing applications. Base-editing technology leverages a catalytically defective Cas9 that is unable to induce DSBs, conjugated with a cytosine deaminase and uracil glycosylase inhibitor or an adenine deaminase. After being guided to the target DNA loci by guide RNA, base editors convert C-G to A-T or A-T to C-G through deaminase.³⁰ By efficiently introducing accurate point mutations, base-editing technology lowers CRISPR/Cas9-associated risks and has been successfully used to construct CD7 CAR-T cells with triple or quadruple gene deletions, showing promising safety and effectiveness.^20,33

Protein expression blocker (PEBL) has also been used to downregulate CD7 surface expression. Instead of ablating the encoding gene, PEBL relies on a single-chain variable fragment (scFv) that targets CD7 and is attached to an intracellular retention amino acid sequence, which anchors CD7 in the ER and/or Golgi apparatus.^54,55 This method efficiently eliminates membrane CD7 protein while retaining the synthesis and expression of CD7 mRNA. Png et al⁵⁴ indicated that PEBL can improve CAR-T cell proliferation and cytotoxicity while mitigating fratricide.

Alternative measures to minimize fratricide without additional genetic manipulations are emerging. For instance, ibrutinib and dasatinib are pharmacologic inhibitors of Itk and Lck/Fyn, respectively, which are pivotal kinases in downstream signaling from CD3ζ.⁵⁶ Inhibition of these kinases can block CAR signaling and prevent fratricide. Watanabe et al⁵⁶ used ibrutinib and dasatinib to temporarily inhibit CAR signaling during CAR-T cell ex vivo proliferation, and removed them before co-culturing with T-ALL cell lines or infusion into mouse models. The results of their study indicated that reversibly inhibiting CAR signaling averted fratricide in CAR-T cells, which fully regained their functionality in vivo upon washout of ibrutinib and dasatinib.⁵⁶

Although the internalization of CD7 upon CD7-targeting-CAR expression was not as obvious as that of CD5 and triggered extensive fratricide formation,⁴⁸ some CAR-T cells survived self-killing. Lu et al¹⁶ exploited the “natural selection” process to generate NSCAR7T cells that were functionally CD7-negative, probably achieved through antigenic masking and/or intracellular sequestration triggered by CD7-targeting CAR. These NSCAR7T cells have exhibited efficacious anti-leukemia activity in clinical trials.¹⁶ Comparatively, naturally occurring CD7-negative T cells in peripheral blood (PB) can also be isolated to generate intrinsically CD7-negative CAR-T cells.¹⁶

3.2. Single-chain variable fragment

As the most crucial accessory that distinguishes CAR-T cells from common T cells, CAR can bind its target antigen with high specificity and affinity in an major histocompatibility complex (MHC)-independent manner through its antigen-binding domain.⁵⁷ Up to date, the most frequently adopted antigen-binding molecule is the scFv derived from murine immunoglobulins, such as TH69 and 3A1e.^{16,17,20,33,48,56} Both TH69 and 3A1e can specifically bind CD7 with high affinity, which paved the way for their successful application in clinical trials.^58–60 Although TH69 exhibited higher affinity than 3A1e in earlier studies, the impact of this difference to CAR-T cell efficacy has not been fully established in clinical trials.⁶⁰ In recent decades, the discovery of camelid heavy-chain only antibodies (HCAbs) has prompted investigations into substituting scFv with nanobodies.⁶¹ Compared with scFv composed of VH and variable light (VL) domains from a monoclonal antibody and a peptide linker, nanobodies consist of heavy-chain variable domains, characterized by their extraordinary solubility, stability, smaller molecular weight, and lower immunogenicity, which are all desired merits for CAR-T cell construction.^62,63 For example, VHH6, a nanobody against CD7, was incorporated into CAR-T cells in a clinical trial conducted by Zhang et al.¹⁴ The novel, autologous, nanobody-derived, and fratricide-resistant CD7 CAR-T cells achieved potent as well as durable anti-leukemia response. Going forward, more research is needed to maximize the superior properties of nanobodies in CAR-T cell production.

3.3. CAR vector

Efficiently expressing the CAR construct in T cells is a pivotal step in CAR-T cell manufacturing, directly affecting the quality and efficacy of the final product. Since frequently used retroviral and lentiviral vectors randomly insert the CAR transgene cassette, both risk insertional carcinogenesis.⁶⁴ For instance, it has been reported that 4 of 9 patients successfully treated by retrovirus-mediated gene therapy developed T cell leukemia.⁶⁵ In contrast, a novel transduction strategy combining CRISPR Cas9 and adeno-associated virus (AAV) can circumvent this safety concern through targeted integration.⁶⁶ Exploiting the homologous-directed-repairing mechanism, Jiang et al incorporated EF1α-driven CD7-specific CAR into the CD7 locus. This targeted 2-in-1 strategy endowed CAR-T cells with enhanced anti-leukemia efficacy, reduced fratricide, and lowered the risk of insertional oncogenesis.⁶⁷ Additionally, the Sleeping Beauty (SB) transposon system, a non-viral delivery strategy, has emerged as an alternative to viral vectors. Compared to virus-mediated CAR transgene delivery, SB transposition offers safer integration, reduced immunogenicity, and lower manufacturing costs.⁶⁸

3.4. Graft-versus-host disease

GVHD occurs when allogeneic immunocompetent CAR-T cells expressing αβTCR recognize the antigen–peptide–MHC complex presented by host cells and are activated, attacking, and damaging normal host organs and tissues. GVHD can be life-threatening, and with the growing interest in developing universal off-the-shelf CAR-T cells, strategies to mitigate or eliminate GVHD are urgently needed.

GVHD essentially appears as a cluster of immune reactions, primarily initiated and mediated by αβTCRs on the surface of T cells in CAR-T products. Consequently, selecting non-αβ T cells to manufacture CAR-modified cell products is a conceivable strategy to ameliorate GVHD. γδT cells, which use a γδTCR for MHC-independent recognition, could be an ideal alternative.⁶⁹ Although γδT cells comprise only around 10% of all T cells, they are actively involved in immune responses and cancer immunosurveillance.^69–71 Notably, δ1T cells, 1 of the 2 subtypes of γδT cells characterized by their superior naïve-like memory phenotype, are particularly noted for their potent cytotoxicity.^70,72,73 Other non-T cell candidates such as NK cells and iNKT cells are also being explored. Like γδT cells, the inherent cytotoxicity of NK cells is not MHC-restricted.⁷⁴ CAR-NK cells have been proven to be efficiently cytotoxic against CD19+ ALL and CD123+ AML cells in preclinical trials.^75,76 Moreover, CAR-NK cells engineered to secrete cytokines like IL-15 have shown improved persistence of NK cells in vivo and in vitro.^77,78 However, the scarcity of γδT and NK cells in PB makes it difficult to generate an ample quantity of CAR-modified γδT and NK cells ex vivo,^28,64,79 a limitation that needs to be addressed before clinical translation can become feasible.

In this regard, induced pluripotent stem cells (iPSCs) stand out for their unlimited self-renewal and multi-lineage differentiation potential.⁸⁰ Various cell types, including γδT cells, can be reprogrammed into iPSCs for specific needs.⁸¹ Theoretically, γδT cell-derived iPSCs are anticipated to provide a bank of substrate cells, which could serve as an unlimited resource for manufacturing allogeneic CAR-T products that circumvent GVHD. Furthermore, iPSCs can differentiate into cells of virtually any specialized type, including γδT cells and NK cells, potentially averting the limitation of in vitro proliferation.^82,83 Similarly, umbilical cord blood (UCB) is another available cell bank where renewable HSCs can be harvested for CAR-T cell production. Van Caeneghem et al⁸⁴ successfully isolated CD34+ HSCs from UCB to produce CAR-T cells with a naïve phenotype and downregulated surface TCR expression.

Studies have shown that removing membrane TCR does not significantly impair CAR-T cell cytotoxicity.⁸⁵ CRISPR/Cas9 remains the most frequently used gene-editing tool for GVHD avoidance because of its relative simplicity and cost-effectiveness. Provided that the β chain contains 2 constant regions encoded by TRBC1 and TRBC2, respectively, knocking out the α chain encoding gene TRAC is always preferred to disrupt αβTCR expression. Multiple trials of CD7 CAR-T cells have confirmed the efficiency of depleting TRAC or TRBC using the CRISPR/Cas9 system to mitigate GVHD.^18,19,29 Combining CRISPR/Cas9 technology with AAV vectors for CAR transgene integration has been shown to be an effective method for synchronously knocking out the TRAC gene locus and inserting the CAR. Essentially, the final CAR-T cells harvested from a single iPSC clone bear identical gene editing and thus are bound to be homogenous, which reduces batch-to-batch variability of the anti-leukemia efficacy of the final product.⁸⁰ When multiple genes beyond TCR are to be edited out, base editing presents as a tolerable tool.^20,33

3.5. Host-versus-graft reaction

HVGR is an immune response similar to GVHD but develops in the reverse direction, in which the host immune system identifies CAR-T cells as foreign and destroys them. Preventing HVGR requires strategies similar to those used for GVHD, with some additional considerations. A primary approach to mitigating HVGR is to abolish surface αβTCR expression, just as is done for GVHD prevention. Additionally, disrupting HLA-I molecules by editing out their common subunit β2 microglobulin (B2M) and silencing HLA-II expression by editing out the encoding gene of their class II transactivator (CIITA) can be an effective strategy.^86–88 Moreover, in clinical cases, intensified lymphodepletion, typically involving fludarabine and cyclophosphamide, is a critical step before administering allogeneic CAR-T cells, with many novel regimens emerging such as anti-CD52 monoclonal antibodies (eg, Alemtuzumab).^20,49 For patients who relapse after HSCT or have an available new HLA-matched donor, HLA-matched donor-derived CAR-T therapy may be an effective approach to reduce the risk of immune rejection and HVGR risk.

3.6. T cell aplasia

T-cell aplasia originates from the on-target off-tumor effects of CAR-T cells that target antigens shared by both normal and tumor T cells. It is the same effect as B cell aplasia, which can however be compensated by periodic infusion of γ-immunoglobins. T-cell aplasia currently lacks a similarly efficacious treatment and predisposes patients to opportunistic infections. Fortunately, various potential solutions are being actively explored to address this challenge.

Selecting more specific target antigens such as CD1a,⁸⁹ TRBC1,⁶⁶ and CD30,⁹⁰ whose expression is limited to a subset of T cells, is thought to prevent T cell aplasia by preserving T cells that are deficient in these antigens. However, CAR-T cells targeting these antigens also exhibit limited applicability, as exemplified by CD30 CAR-T cells for classical Hodgkin lymphoma,⁹¹ and CD1a for cortical T-ALL.⁹² Utilizing short-lived cell types, including NK cells, NK92 cells, or γδT cells, to construct CAR cell products is advantageous in lowering the risk of inducing T cell aplasia, which has been correlated to the long-term existence of CAR-T cells. The absence of memory cell formation also contributes to circumventing persistent normal T cell elimination by CAR-modified cells. Additionally, using non-integrating delivery methods, like AAV vectors or mRNA electroporation, can result in transient cytotoxic activity of the CAR-T cells compared to transduction via retroviruses or lentiviruses. After several rounds of mitosis, AAV expression is substantially diluted, resulting in a CAR expression density below the effective level.⁹³ Similarly, as the translation of mRNA is a one-shot process, CAR expression diminishes as the mRNA level decreases. When patients eventually develop T-cell aplasia, termination of its progression is more potent and emergent.

Exploiting HSCT, especially with myeloablative preconditioning, can be used to clear CAR-T cells and break T-cell aplasia. Nonetheless, availability and HSCT-associated AEs should be considered. Incorporating suicide genes by expressing them with CAR in a bicistronic vector enables the efficient elimination of CAR-T cells. The inducible caspase-9-based suicide gene (iCas9) system rapidly eradicates CAR-T cells upon administration of a chemical inducer of dimerization (CID).⁹⁴ The antibody-dependent cellular cytotoxicity (ADCC) mechanism requires the introduction of CD molecules that can be targeted by available antibodies, such as CD19 and CD52.

4. FUTURE PERSPECTIVE

CAR-T therapy has transformed blood cancer treatment with the success of CD19 CAR-T treatment spurring its broader use. This progress has catalyzed the exploration of CD7 CAR-T products, showing promising short-term CR in clinical trials. Nevertheless, long-term anti-leukemia potency, which is key to relapse prevention, is a roadblock for optimizing CD7 CAR-T cells. CAR-T cell persistence may cause serious AEs such as prolonged T-cell aplasia and immunodeficiency, an increased risk of opportunistic infection, and virus reactivation. Researchers are actively working on strategies to balance leukemia control with AE prevention. The development of controllable CAR-T cells is currently being vigorously investigated. One strategy is to incorporate suicide genes and safety switches into the CAR construct, including iCas9, EGFR, and CD20. The iCas9 system, in particular, offers the ability to rapidly and selectively eliminate transduced T cells with minimal immunogenicity. Preclinical trials have shown that iCas9 can eradicate CAR-T cells in a dose-dependent manner upon administration of a CID, demonstrating its flexibility in regulating the specific containment of CAR-T cells. Combinatorial therapy involving CAR-T cell transfusion with sequential HSCT represents a novel trend toward maximizing the anti-leukemia potency of cellular immunotherapy. Patients with r/r T-ALL/LBL are predominantly chemo-resistant and thus unable to achieve CR, which is a prerequisite for HSCT. Given this background, CAR-T-cell therapy has blazed a new trail in offering innovative treatment options for these patients. HSCT has been tentatively used as a consolidative regimen for CAR-T treatment. However, previous clinical trials included routine myeloablative conditioning and GVHD prophylaxis, which impaired the efficacy of CAR-T cells.^17,95 Hu et al⁴⁶ proposed an innovative combinatorial regimen. In their cohort of 10 patients with r/r T-ALL/LBL, the participants were treated with sequential allogeneic CD7 CAR-T cell therapy and haploidentical HSCT, discarding myeloablative and GVHD prevention treatment. Surprisingly, 60% of patients remained in MRD-negative CR until the data cutoff date, with a 15.1-month median follow-up period after CAR-T cell infusion, achieving survival data comparable to those of younger patients who underwent HSCT after CAR-T cell therapy. Eight of 10 patients achieved full donor chimerism, which could be attributed to lymphodepletion pretreatment as well as myelosuppression and pancytopenia induced by CAR-T therapy. Only 3 developed low-grade acute GVHD, which may have resulted from depletion of CD7-positive alloreactive T cells mediated by persisting CAR-T cells.⁴⁶ In brief, CAR-T therapy’s progress highlights areas for refinement before seamless application in clinical practice.

5. CONCLUSION

Universal CD7 CAR-T cells are expected to play a dominant role in the treatment of T-cell malignancies and some forms of AML. Sophisticated gene-editing technologies and post-translational manipulations hold great potential for overcoming existing difficulties, such as fratricide, GVHD, and HVGR. Additionally, iPSCs are a potentially revolutionary source of T cells for future universal CAR-T cell manufacturing. The rapid development of CAR-T therapy is bound to be used in a new era of adoptive cellular immunotherapy.

ACKNOWLEDGMENTS

Funded by Tianjin Municipal Science and Technology Commission Grant (23JCYBJC01050), CAMS Innovation Fund for Medical Sciences (2024-I2M-TS-023, 2024-I2M-ZH-015, 2021-I2M-1-041).