Can less differentiation drive CARs to success?

In this issue of Blood, Caballero et al¹ present the results of a small phase 1 clinical trial of second-generation CD30-directed chimeric antigen receptor T cells (CAR-T). This product (HSP-CAR30) was manufactured with the goal of promoting less differentiated memory T cells and used to treat patients with CD30⁺ lymphoma, with half of patients achieving durable remissions.

CD30 is an ideal target for CAR-T therapy given its expression in classical Hodgkin lymphoma (HL) as well as other lymphomas, including T-cell lymphoma, and limited expression in normal tissues.² To date, several clinical trials investigating CD30-directed CAR-T (CD30.CAR) have been published, showing some promising responses.^3-6 In the largest published study, Ramos et al infused 41 patients with heavily pretreated relapsed or refractory HL with a CD30.CAR containing the CD28 costimulatory domain.⁴ The overall response rate was 72% with a complete response (CR) rate of approximately 60%. Durable responses were observed in some patients, with 5 remaining in CR longer than 1 year post infusion. However, the majority of patients eventually relapsed, with a 1-year progression-free survival (PFS) of 41%.

In the current study, 10 patients (8 with HL and 2 with T-cell lymphoma) were treated with HSP-CAR30. Similar to prior CD30.CAR trials, these patients were heavily pretreated, with most refractory to brentuximab vedotin and PD1 inhibitors. For lymphodepletion this trial used fludarabine/bendamustine (though at a higher dose than in previous studies⁴), confirming this is a valid regimen prior to CAR-T infusion in patients with HL. The toxicity profile of HSP-CAR30 was comparable to other CD30.CAR studies, with only low-grade cytokine release syndrome and no neurotoxicity. Skin rash, a recurrent feature of CD30.CAR studies, occurred in 4 patients, resolving without therapy in all but 1 patient, who required tropical steroids. The rates of cytopenias remained comparable to other CAR-T trials, although slightly higher rates of infection (primarily viral) were observed. All patients responded, with 5 patients achieving CR (all with HL) and an additional patient obtaining a CR after a second infusion of CAR-Ts. At a median follow-up of 34 months, 60% of patients were in ongoing CR, with a 2-year PFS of 40%, suggesting the potential for more durable remissions with HSP-CAR30.

What sets HSP-CAR30 apart from the other studies is the design and manufacturing differences. Firstly, it has a 4-1BB costimulatory domain, and most prior studies’ CD30.CAR constructs used a CD28 costimulatory domain. In other constructs (like CD19.CAR), 4-1BB costimulation has often been associated with improved persistence compared with CD28, which seems to imprint T cells with more rapid antitumor activity.⁷ Given that Hodgkin Reed-Sternberg cells constitute a minority of the tumor environment,⁸ prioritizing prolonged persistence over rapid activity may be advantageous. A second difference is in the targeted epitope. HSP-CAR30 targets a proximal, noncleavable region of CD30 to mitigate the risk of CAR-T inhibition by the soluble CD30 protein.⁹ Distal epitopes may be susceptible to inhibition by high levels of soluble CD30, although no inhibition has been demonstrated to date in ex vivo and in vivo studies.¹⁰ 

The most critical difference, however, lies in the manufacturing strategies. Caballero et al focused on increasing the proportion of less differentiated memory T cells in the infusion product. This was achieved by incorporating interleukin-7 (IL-7), IL-15, and IL-21 cytokines during manufacturing, to enrich memory stem and central memory T cells. Additionally, the use of the CliniMACS Prodigy platform enabled a shorter manufacturing process that avoided cryopreservation, resulting in an impressive vein-to-vein time of just 11 days. The final product demonstrated a less differentiated T-cell phenotype and molecular signature, characterized by high expression of genes associated with T-cell stemness and memory. Functionally, this resulted in a higher number of early memory CAR30⁺ T cells at peak CAR-T expansion and long-term persistence, with CAR30⁺ T cells detected in 3 of 5 patients at 1 year post infusion.

The shorter production time offers several advantages, including reduced need for bridging therapies and lower risk of disease progression during manufacturing. However, the infusion of fresh products introduces logistical challenges, particularly scalability and timing of lymphodepletion. Initiation of lymphodepletion before completing manufacturing risks leaving patients exposed to immunosuppressive lymphodepletion without CAR-T infusion. Fortunately, only 1 patient in this study had a production failure. Larger studies will be needed to assess the feasibility of this approach on a large scale. The use of the CliniMACS Prodigy system may facilitate such studies, enabling a decentralized process and thus improving accessibility. However, a centralized quality control system may still be necessary to harmonize products across centers.

Although the results of this phase 1 trial are encouraging, particularly in HL, longer follow-up and a larger sample size are needed to determine whether HSP-CAR30 promotes more durable remissions compared with prior CD30.CAR products. Mechanisms of resistance of CD30.CAR remain an open question. Most patients relapsing post-CD30 CAR-T retain CD30 expression in their tumor,^4,6 suggesting that antigen escape is not a predominant mechanism of resistance. The superior performance of these less differentiated cells may stem from additional inherent properties. Understanding these mechanisms will be critical for developing strategies to extend remissions. Future studies should also explore combination therapies (eg, with PD1 inhibitors) or multiple infusions. Consolidation strategies, perhaps with the integration of biomarkers such as circulating tumor DNA, may help refine therapeutic approaches. Still, the optimization of the infusion product remains key.

Finally, this study reinforces the need for more effective therapies for patients with T-cell lymphoma. In this trial, neither of the 2 treated patients with T-cell lymphoma achieved a CR, and both passed away less than 1 year after treatment. This high-risk population has consistently posed challenges, with inferior outcomes. Even with a more optimized manufacturing process, additional therapeutic strategies may be required to improve outcomes in T-cell lymphoma.

Conflict-of-interest disclosure: N.S.G. has received research funding from Cabaletta, Bristol Myers Squibb, Affimed, Regeneron, and Poseida, and has received honoraria from Novartis, Bristol Myers Squibb, ADC Therapeutics, Regeneron, and Genentech. B.S. declares no competing financial interests.