Educational theme | Tumor intrinsic resistance mechanisms to CAR T-cell therapy in lymphomas

Chimeric antigen receptor (CAR) T-cell therapies have demonstrated significant efficacy in patients with relapsed/refractory B-cell malignancies following multiple lines of therapy, with various CAR T cell products now U.S. Food and Drug Administration (FDA) approved for the treatment of B-cell acute lymphoblastic leukemia (B-ALL), diffuse large B-cell lymphoma (DLBCL), mantle-cell lymphoma (MCL), and follicular lymphoma (FL).¹

Despite this, longer-term survival outcomes (5-year PFS rates of 31% and 43% for patients with DLBCL and FL after tisa-cel (tisagenlecleucel) treatment, respectively)² show that the majority of patients will experience disease progression or resistance to CAR T-cell therapy. As such, there is a need to elucidate the underlying mechanisms of treatment resistance in lymphoma, which can include CAR T-cell dysfunction, immunosuppressive tumor microenvironment, and tumor-intrinsic factors.^1,2

The Lymphoma Hub has previously reported on the key genomic features underlying resistance to CD19-directed CAR T-cells. Here, we focus on the tumor intrinsic mechanisms of resistance to CAR T-cell therapy in patients with lymphomas, discuss novel strategies for managing treatment resistance using data published by Lemoine et al. in Journal of Hematology and Oncology¹, and summarize data presented by Ruella at the recent EBMT-EHA 5th European CAR T-cell Meeting.²

Tumor intrinsic mechanisms

Tumor intrinsic mechanisms to CAR T-cell therapies include loss of the target antigen, expression of inhibitory receptors, lack of costimulatory ligands, and resistance to immune killing.¹

Loss of target antigen^1,2

While the loss of target antigen is well documented after CAR T-cell therapy relapse in B-ALL, similar data in lymphomas is lacking. However, loss of CD19 and CD20 and retention of CD79a has been observed in lymphoma biopsies by immunohistochemistry in tisa-cel treated patients experiencing relapse^1,2; a 30% incidence of CD19 loss has also been reported in patients with lymphoma relapsing following axi-cel (axicabtagene ciloleucel) treatment.²

Mechanisms of antigen-negative escape to CD19-directed CAR T-cell therapies, which have been largely studied in B-ALL, include convergence of acquired mutations; alternative splicing of CD19, which generates CD19 isoforms with disruption of the target epitope; and transdifferentiation/lineage switching, including the change from lymphoid to myeloid blast. The basis for other target antigen loss can include selection of pre-existing antigen-negative tumor cells, altered maturation/trafficking affecting target antigen expression, and epitope masking (Figure 1).^1,2

 Strategies to overcome CD19-negative escape include pooled CAR T cells, dual-targeted CAR T cells, tandem and 4th generation CAR T cells, and combination CD19-directed CAR T cell-therapy (CD19 CAR T-cell) and bispecific antibodies.²

Given the retention of CD79 antigen on lymphoma cells after CAR T-cell relapse, it was considered a suitable therapeutic target for dual CAR T-cell therapy. Among various anti-CD79b CAR T-cell constructs, the CART79#3 structure performed best in murine models when compared with CD19 CAR T-cell candidates. The dual anti-CD19/79b CAR T-cell construct demonstrated tumor cell killing on double-positive blasts compared with the single anti-CD19 and anti-CD79 CAR T-cell therapies. In CD19 and CD79b knockout mice model, this dual CAR T-cell therapy also proved more favorable when compared with single anti-CD19 CAR T-cell and control agents; this is now being explored in clinical trials.²

The combination of CD19 CAR T cells and mosunetuzumab in non-Hodgkin lymphoma is another novel strategy being investigated to overcome resistance in a phase II trial (NCT04889716), with a two-fold concept in mind, 1) targeting both CD19 and CD20 and 2) restimulating CAR T-cell activity by administering mosunetuzumab 30 days after the CAR T-cell infusion.²

Apoptosis²

Apoptosis is a key tumor-intrinsic resistance mechanism of CD19 CAR T cells. A genome-wide CRISPR-based loss-of-function screen, targeting over 20,000 genes in tumor cells at relapse, detected enrichment of pro-apoptotic genes and depletion of anti-apoptotic genes; this indicates the involvement of impaired apoptotic signaling in tumor resistance.

Additionally, a small molecule screen of over 3,000 molecules, alongside CAR T-cells in tumor cells, found that SMAC mimetics (birinapant) and BCL-2 inhibitors (venetoclax) enhanced anti-tumor activity of CAR T-cell therapy. Whilst birinapant proved efficacious within in vitro studies, it was conversely shown to induce apoptosis of CAR T-cells in vivo.

Among a cohort of 87 patients with large B-cell lymphoma treated with CD19 CAR T-cells, those who harbored BCL-2 alterations (either translocation or gain-of-function) yielded lower overall response and overall survival (OS) rates compared to those without BCL2-alterations; this suggests that BCL2 inhibition could be a beneficial treatment strategy for this patients subgroup.

In xenograft mouse models of B-ALL and MCL, venetoclax combined with CD19 CAR T cells initially demonstrated antitumor effects; however, its longer-term effects were worse compared with the CD19 CAR T-cell therapy alone. Peripheral blood analysis showed lower CAR T cell levels after combined venetoclax plus CD19 CAR T cells compared with CD19 CAR T-cell monotherapy treatment, indicating that BCL2-inhibition affects CAR T-cell proliferation and likely activates its cell-induced death.

To optimize the therapeutic use of CAR T cells with venetoclax, the F104L mutation, known to confer resistance to venetoclax in tumor cells, was used for development of a venetoclax-resistant CAR T-cell therapy (CART19-CL-2[F104L]). This construct was shown to resist venetoclax and protect CAR T- cells in vitro and was synergistic when combined with venetoclax in vivo; this resulted in 100% OS. Although there is limited data on this combination in clinical settings, a retrospective analysis of venetoclax as bridging therapy prior to brexu-cel (brexucabtagene autoleucel) infusion in patients with MCL demonstrated a higher complete response (CR) and event-free survival rate compared with patients who did not receive venetoclax-based bridging therapy.

Overexpression of wild-type BCL-2 in CD19 CAR T-cells is another investigative treatment strategy shown to enhance its in vivo antitumor activity, with prolonged persistence and peak expansion of CAR T-cells also observed. Among patient T cells analyzed from apheresis products after CAR T-cell therapy, higher BCL-2 expression was seen among those achieving a complete vs no response; this also correlated with prolonged CAR T-cell persistence and OS.

Expression of inhibitory ligands¹

Another tumor intrinsic mechanism of resistance to CAR T-cell therapy can involve expression of inhibitory ligands, programmed death-1 ligand-1 (PD-L1) and programmed death-2 ligand-2 (PD-L2); both may prevent CAR T-cell activation and induce immune resistance. In a ZUMA-1 trial (NCT02348216) analysis, PD-L1 was expressed in 62% (13/21) of patients with DLBCL who progressed after axi-cel therapy.

Strategies to overcome PDL1-mediated resistance include combining CAR T cells with anti-PD1/PD-L1 antibodies, anti-PD1-secreting CAR T-cell therapy, and PD-L1-resistant CAR T-cell therapy. A single-center study, including 12 patients with DLBCL who relapsed after CD19 CAR T-cell therapy, investigated the efficacy of sequentially adding the PD1 inhibitor pembrolizumab; of 11 evaluable patients, one achieved a CR, two a partial response, one achieved stable disease, and seven patients experienced disease progression. An overall response rate of 50% was reported in a small study of patients with DLBCL (n = 15) treated with the PD-L1 inhibitor durvalumab, either 21–28 days after (n = 6) or 1 day prior to (n = 9) CAR T-cell infusion, with only 1/5 patients experiencing disease progression after CR.

Overall, 9/10 patients with relapsed/refractory DLBCL achieved an objective response (six with a CR and 3 with a PR) in a ZUMA-6 trial (NCT02926833) investigation the anti-PD-L1 antibody atezolizumab administered subsequent to CD19 CAR T-cell infusion. Engineered CAR T cells with resistance to PD1-signaling are shown to resist inhibition within preclinical models of B-cell lymphomas; however, clinical data are not yet available.

Lack of costimulatory ligand¹

Mutations or loss of expression of CD58, a costimulatory ligand for the CD2 molecule involved in T-cell proliferation, is a tumor intrinsic mechanism which is shown to decrease CAR T-cell activation and cytotoxicity within preclinical models. This accounts for approximately 25% of relapsed cases following CAR T-cell therapy in patients with DLBCL and has been associated with poorer outcomes.

The addition of a CAR construct containing a CD2-signaling domain in the cytoplasmic tail could be a potential novel strategy to overcome resistance. So far, it has increased activity in preclinical models of large B-cell lymphomas, but its clinical activity in patients has not yet been established.

Conclusion

Understanding the mechanisms of tumor intrinsic resistance to CAR T-cell therapies is fundamental for the development of novel strategies. Various strategies are being explored to tackle the loss of target antigen, including pooled and dual CAR T cells that incorporate alternative target antigens and investigative combination therapies such as CAR T cells and bispecific antibodies. Strategies have also been devised to ameliorate the impact of impaired apoptotic pathways, including BCL2-inhibition resistant CAR T-cell therapies and BCL2 overexpression. Moreover, CAR T cell/immunotherapy combinations and re-engineered CAR constructs, such as those with a CD2 signaling domain and anti-PD1 secreting CAR T cells, are being investigated to address CD58 loss and PDL1 overexpression in tumor cells. Future work within this area is likely to further optimize the efficacy and personalized use of CAR T-cell therapy in lymphomas.