Subverting the immunoproteasome in acute myeloid leukemia

Disease relapse continues to be a roadblock in the successful treatment of acute myeloid leukemia (AML). To improve relapse risk and prolong remission, tremendous effort has focused on T-cell directed therapies. T-cell receptor (TCR) therapy involves engineering T-cells to bind to peptides present on the surface of cancer cells but not healthy cells, for a precise, targeted treatment. Miranda Lahman, Dr Thomas Schmitt, Dr. Kelly Paulson, Dr. Nathalie Vigneron, Dr. Denise Buenrostro, Dr. Aude Chapuis, and Dr. Phil Greenberg, all members of Fred Hutch’s Clinical Research Division, and their colleagues have focused their recent research efforts on generating TCRs that recognize peptides associated with Wilms’ tumor antigen 1 (WT1), a factor highly expressed on AML tumor cells. Previous research on WT1 had great success in the clinic, and a WT1 TCR previously generated by the authors, termed T_TCR-4, demonstrated promising success in patients with AML. Specifically, the engineered T-cells remained present in circulation and prolonged remission in patients. In their current study, recently published in Science Translational Medicine, they made an interesting observation when tracking the progress of a patient who relapsed after T_TCR-4  treatment. The patient had initially responded well to treatment, with a remission much longer than anticipated (almost four times longer than expected) and still had circulating T cells which prompted the authors to question why WT1+ AML cells had stopped responding to treatment, a question which eventually led them to the immunoproteasome. Their study finds that T_TCR-4 cells do not adequately clear WT1+ tumor cells with limited immunoproteasome subunit expression and identifies an additional TCR (T_TCR37-45) that is not dependent on the immunoproteasome with potential to circumvent the described relapse and to treat solid tumors.

With the goal of better understanding this patient’s remission and subsequent relapse, the authors took a closer look at blood samples throughout therapy. They observed that T_TCR-4 cells remained persistent in the blood throughout remission and relapse. Using single-cell RNA-sequencing the authors looked at the transcriptome of both T_TCR-4 cells and AML at remission and relapse. Initial gene expression analysis determined that genes associated with T-cell activation were differentially expressed during remission when compared to relapse, confirming transcriptional differences at the two time points. After establishing that WT1 expression was not lost or mutated during relapse, the authors compared AML to non-AML hematopoietic cells, finding a significant reduction in immunoproteasome subunit β1i in the AML cells. Exploring further, they noted that after the TCR-T therapy, the AML had a decreased level of transcripts encoding for β1i, which restricts development of the immunoproteasome; this decrease was observed at both the protein level and transcript level (gene: PSMB9). Using in vitro cell line models, they examined whether T_TCR-4 treatment was dependent on immunoproteasome expression and found that T_TCR-4 T-cells only recognized cells expressing the β1i subunit. “We further build on the existing knowledge that the peptide targeted by T_TCR-4 is immunoproteasome dependent by demonstrating that it is dependent on the particular immunoproteasome subunit, β1i (gene: PSMB9),” explained Lahman.

In a second patient who relapsed, the authors were able to confirm this dependency on the immunoproteasome by confirming the T_TCR-4 therapy only reduced WT1+ AML cells that also co-expressed PSMB9, but not the WT1+ AML cells without PSMB9 expression. To address this problem, they focused on generating additional engineered T-cells that were not immunoproteasome dependent, in the hope that this could reduce relapse risk for patients. Alongside Ruggiero et al., they identified a new TCR, T_TCR37-45, recognizing another portion of WT1, that demonstrated promising effects in vitro by successfully targeting WT1-expressing AML cells. Additional experiments confirm that T_TCR37-45 is not dependent on the immunoproteasome. Lastly, they tested the efficacy of T_TCR37-45 in vivo by utilizing mouse models engrafted with a WT1⁺β1i^- solid tumor line. They observed T_TCR37-45 was more successful at targeting the WT1⁺β1i^- solid tumor than T_TCR-4, thus highlighting the potential as a non-immunoproteasome dependent therapy. Summarizing their study findings, Lahman said “we identified and validated a new TCR targeting a peptide that is not reliant on the immunoproteasome and show potential for use of this TCR in solid tumors which are typically immunoproteasome negative.”

The Chapuis lab and colleagues are continuing to focus their research efforts on the development of additional TCRs, aiming to reduce the risk of relapse for patients with AML. Discussing their ongoing research, Lahman stated “this work specifically highlights the need to understand and validate the proteasomal processing of TCR target peptides early in the development pipeline. By identifying peptides that are generated by all proteasomal isoforms we aim to minimize escape through shifts in proteasome composition as observed in this case study. We are partnered with the Strong Lab to identify such ‘mutual peptides’ (stay tuned!). For promising TCR targets already identified, we have developed assays to determine if the candidate peptide has proteasome dependencies like T_TCR-C4. We have already used this assay within our lab and in collaborations with others at the Hutch.  We are also curious how frequently cancer uses this avenue of immune-evasion not only in the context of TCR-T cell immunotherapy but also within the cancer-immune interface.” She continued by acknowledging the role of clinical teams in the success of this study “this work would not be feasible without the patients nor without the doctors, nurses, clinical coordinators, and staff at the multiple institutions of the Cancer Consortium that make these complex clinical trials possible.”

This work was supported by funds from the National Institutes of Health, SITC-Merck Immunotherapy Fellowship, Immunotherapy Integrated Research Center at Fred Hutch, Damon Runyon, Guillot Family Zach Attacks Leukemia Foundation, Juno Therapeutics, and the Ludwig Institute for Cancer Research.

UW/Fred Hutch Cancer Consortium members Jason Bielas, Julie M. McElrath, Derek L. Stirewalt, Cecelia C. Yeung, Paul C. Hendrie, Marie Bleakley, Anthony Rongvaux, Aude G. Chapuis and Philip D. Greenberg contributed to this work.

Lahman MC, Schmitt TM, Paulson KG, Vigneron N, Buenrostro D, Wagener FD, Voillet V, Martin L, Gottardo R, Bielas J, McElrath JM, Stirewalt DL, Pogosova-Agadjanyan EL, Yeung CC, Pierce RH, Egan DN, Bar M, Hendrie PC, Kinsella S, Vakil A, Butler J, Chaffee M, Linton J, McAfee MS, Hunter DS, Bleakley M, Rongvaux A, Van den Eynde BJ, Chapuis AG, Greenberg PD. Targeting an alternate Wilms' tumor antigen 1 peptide bypasses immunoproteasome dependency. Sci Transl Med. 2022 Feb 9;14(631):eabg8070. doi: 10.1126/scitranslmed.abg8070. Epub 2022 Feb 9. PMID: 35138909; PMCID: PMC8985599.