AUTHORS
Kirsten Pfeffer, PhD, Yuan Xiao Zhu, PhD2, Chang-Xin Shi, PhD3, Meaghen Sharik, BS4, Megan T. Du, MS5, Yuliza Tafoya Alvarado, MS6, Lorenzo Lindo, BSc7, Kevin Hay, MD, MSc, FRCPC8, Erin Meermeier, PhD9, P. Leif Bergsagel, MD10, Marta Chesi, PhD11
AFFILIATIONS
1 Mayo Clinic, Scottsdale, Arizona, United States
2 Mayo Clinic, Scottsdale, Arizona, United States
3 Mayo Clinic, Scottsdale, Arizona, United States
4 Mayo Clinic, Scottsdale, Arizona, United States
5 Mayo Clinic, Scottsdale, Arizona, United States
6 Mayo Clinic, Scottsdale, Arizona, United States
7 BC Cancer Research Institute, Vancouver, British Columbia, Canada, University of British Columbia, Vancouver, British Columbia, Canada, University of Calgary, Calgary, Alberta, Canada
8 BC Cancer Research Institute, Vancouver, British Columbia, Canada, University of British Columbia, Vancouver, British Columbia, Canada, University of Calgary, Calgary, Alberta, Canada
9 Mayo Clinic, Scottsdale, Arizona, United States, Mayo Clinic, Scottsdale, Arizona, United States
10 Mayo Clinic, Scottsdale, Arizona, United States
11 Mayo Clinic, Scottsdale, Arizona, United States
ABSTRACT
Introduction
Despite high response rates, a large fraction of multiple myeloma (MM) patients relapse to BCMA CAR-T cell therapy. Clinical studies have shown that patients with a low disease burden at CAR-T infusion have better PFS and favorable safety profiles. In contrast, high tumor burden (HTB) and poor CAR-T persistence are associated with inferior responses and greater risk for toxicities, so bridging therapy is recommended to reduce tumor burden prior to CAR-T infusion. While there is a correlation between tumor burden and outcome, controlled preclinical studies are needed to adequately assess CAR-T response in HTB and define the mechanism by which tumor burden impairs CAR-T function.
Having demonstrated in vitro effects of tumor burden on the activity of murine BCMA CAR-T cells, we used the immunocompetent Vk*MYC model of MM to characterize CAR-T dysfunction in vivo that contributes to HTB primary resistance.
Methods
Murine anti-BCMA.CD28 CAR-T cells were generated and tested in vitro and in vivo low and high tumor burdens using the Vk*MYC line, Vk12653. 1x106 (low dose) or 5x106 (high dose) CAR-T cells were infused at 1-week (LTB) or 4 weeks (HTB) post-transplant in lymphodepleted (LD, cyclophosphamide + fludarabine) C57BL/6J or RAGnull mice. Tumor burden was monitored by M-spike and CAR-T cells from blood, bone marrow and spleen were profiled ex vivo by flow cytometry.
Results
CAR-T cytotoxicity was significantly impaired by HTB. In vitro, HTB (low E:T) induced equivalent levels of CAR-T activation and expansion compared to LTB but drove early CAR-T apoptosis and death resulting in decreased CAR-T cells and reduced killing. Resistance was not overcome by rIL-7 or PD-1 blockade. As expected, addition of IMiD (POM) boosted CAR-T expansion in LTB and was surprisingly deleterious in HTB by promoting CAR-T cell death and exhaustion.
In vivo, CAR-T cells were universally curative when infused at LTB in LD tumor bearing mice. In contrast, only a high dose of CAR-T in HTB demonstrated transient tumor control, and mice inevitably relapsed. While no differences between LTB and HTB setting were noted in the phenotype of CAR-T collected day 3 post-infusion, analysis on day 10 showed decreased absolute CAR-T number in HTB compared to LTB mice. This was coupled with increased apoptosis, reduced proliferation, cytotoxic activity, and effector phenotype compared to LTB. To improve responses in HTB, we pretreated tumor bearing mice with a higher dose of cyclophosphamide (Cy300), which deepened LD but also reduced tumor burden, resulting in improved CAR-T expansion and function, and curative responses. Finally, to distinguish tumor debulking from LD effects of Cy300, we infused CAR-T cells in HTB RAGnull mice, where we noted transient tumor control but not cure, suggesting that tumor burden, more than extent of LD dictates CAR-T efficacy. Interestingly, in this setting, treatment with POM transiently augmented hCRBN+ CAR-T expansion and deepened the response without promoting CAR-T longevity.
Conclusion
Although studies link MM disease burden with early relapse, there is a need to evaluate CAR-T cell changes in HTB that contribute to treatment failure. We previously demonstrated that primary resistance to BCMA/CD3 T cell engagers is driven by tumor burden and induces T cell dysfunction. Likewise, response to CAR-T is dictated by tumor burden as CAR-T is curative in LTB but have limited efficacy in HTB due to CAR-T dysfunction.
Without LD, CAR-T cannot survive due to physical constraints and metabolic competition within the TME and T cell niches. With LD, CAR-T expand in LTB then contract rapidly, attributable to successful tumor clearance. In contrast, HTB pushes CAR-T into apoptotic cell death, decreases cytotoxicity and limits expansion, reflecting an early functional collapse that hinders response. To overcome this, high dose LD was used to debulk the tumor and promote CAR-T engraftment to improve expansion and response.
Efficacy requires bridging therapy for tumor reduction and LD for adequate expansion of CAR-T that favor a proliferative, cytotoxic state. For patients that do not respond to bridging, there is a need for alternative tumor debulking strategies and ways to enhance and preserve CAR-T function. Future studies will evaluate how HTB alters CAR-T metabolism, and test IMiDs to boost early effector CAR-T function that accelerates tumor reduction, with subsequent therapies aimed to preserve memory phenotypes for persistence.
DISCLOSURES
Kirsten Pfeffer, PhD: No relevant conflicts of interest to declare.