Cancer Immunotherapy: Adoptive Cell Therapies and Beyond

Keynote: Unraveling Mechanisms of Response and Resistance to Immune Checkpoint Therapy

The concept of yin and yang aptly describes the immune response, wherein activating signals drive inhibitory feedback within the same cell or larger environment. Physician scientist Padmanee Sharma pays special attention to this duality as she interrogates mechanisms of response and resistance to cancer immunotherapy. Sharma and her team developed Reverse Translation, a platform to address key questions in the field. The platform takes advantage of patient observations across thousands of clinical trials to generate hypotheses, which are then tested in the laboratory setting. Uncovered mechanisms may then inform additional treatment methodologies towards benefiting patients in subsequent or revised clinical trials.

The Reverse Translation platform enables rapid testing of hypotheses generated by patient observations.

Using Reverse Translation, Sharma’s team investigated the effects of anti-CTLA-4 on immune response within the tumor microenvironment (TME) of bladder, melanoma, prostate, and glioblastoma patients, unearthing actionable targets to improve therapeutic regiments in each case. Sharma highlighted findings in bladder cancer patients, who currently don’t have any standard of care treatments prior to surgery. Comparing differentially expressed genes in pre- and post-treated biopsies, a significant increase of inducible co-stimulator (ICOS)+ T cells was observed in treated samples. Patients who sustained elevated levels of ICOS+ T-cells had markedly improved survival outcomes compared to those who did not. Responders also exhibited a higher density of tertiary lymphoid structures within the TME, indicating that ICOS may play a role in their formation.

Combining CTLA-4 blockade with an ICOS agonist led to remarkable responses and long-term survival in ICOS wild-type mice compared to their knockout counterparts. “The combination therapy can really lead to important clinical responses that hopefully will [prompt] a new standard of care therapy for these patients with cisplatin-ineligible disease,” Sharma concluded.

What COVID-19 and Driver Genes Teach us about Quantifying Immune Recognition

Benjamin Greenbaum applies computational methods to quantify the interaction between tumors and the immune system and predict immune-driven tumor evolution. Part of Greenbaum’s research aims to achieve specificity in quantifying immune recognition of neoantigens by leveraging various datasets. To this end, Greenbaum focused on p53 hotspot mutations. “Part of what makes you a hotspot mutation for a driver gene is that you’ve escaped the immune system,” Greenbaum began. “On average, the peptides generated by driver gene mutations tend to present a bit more poorly.”

Greenbaum’s team generated a p53 quantitative fitness model that considered a given mutations’ affinity for loss of functional capacity to bind DNA (oncogenicity), and its capacity to be discovered by the adaptive immune system (immunogenicity). The model demonstrated that increasing loss of function (LOF) precipitated by a mutation leads to greater immunogenicity since LOF increases p53 concentration, making it a better neoantigen. Hence, a tumor cell must balance oncogenicity with immune system evasion, with increased LOF and reduced immunogenicity yielding a great hotspot mutation.

TGFβ in Cancer Immunology and Immunotherapy

Shannon Turley studies patient responses to cancer immunotherapy (CIT) to leverage findings to convert non-responders to responders, and to improve the depth and duration of treatment in responders. Turley and her team evaluated pathways associated with response, or lack thereof, to anti-PD-L1 in bladder cancer patients. Bladder cancer is an immune-excluded tumor, characterized by plentiful T cells in the tumor microenvironment (TME) that cannot infiltrate the tumor well. Turley’s team found that TGFβ pathway-related genes were enriched in the non-responder setting.

Combining TGFβ inhibition with anti-PD-L1 promoted response and survival in the EMT6 breast cancer model, yielding roughly 70% CD8 T cell-dependent tumor regression.  “What is striking is that every single cellular compartment of the tumor is changing in response to this combination therapy,” Turley stated. Among these changes were reduced proliferation of immunosuppressive macrophages in the myeloid compartment, a dampening of fibro-progenitor programs in the fibroblast compartment, and an augmentation of inflammatory programs in the tumor cell compartment. Hence, the combination converted the landscape of the TME from immune excluded to immune inflamed.

Keynote: Normal and Neoplastic Stem Cells

Hematopoietic stem cells (HSCs) are a rare type of blood cell (1 in 100,000) responsible for generating the entire blood-forming network. As such, errors within these cells may result in malignant transformation, adult onset blood diseases—such as myelodysplastic syndrome (MSD)—and other disorders. Irving Weissman is widely accepted as the “father of hematopoiesis” for his seminal work in purifying and characterizing HSCs from mice and humans. Weissman’s lab also spent considerable time defining the hematopoietic hierarchy and locating where the cell of origin for leukemia in its various forms arose in this system. This research led to the discovery of the leukemic stem cell (LSC) for acute myelogenous leukemia (AML), arising from the multipotent progenitor (MPP) population, and for Chronic Myelogenous Leukemia (CML), stemming from the granulocyte monocyte progenitor (GMP) population. “Understanding stem cells is important for moving the field,” said Weissman. “Only if the mutation occurs in a self-renewing cell—which in the HSCs is just that 1 in 100,000 cells—could a clone be built that gains all of the events that give rise to the LSC.”

Oncogenic mutations and translocations occur with equal frequency in all cells, but will only give rise to leukemia if the cell is naturally self-renewing, or gains self-renewal capacity.

Weissman’s lab observed increased expression of CD47, a ‘don’t eat me’ signal to the SIRPα receptor on macrophages, in AML LSCs. They found that activated macrophages secrete calreticulin (calR), the dominant cell surface ‘eat me’ signal, for Programmed Cell Death Removal (PrCR). calR binding to target cell asioglycans promotes phagocytic removal of cancer, dying, and pathogenic cells. Emergent cancer cells overcome calR by upregulating CD47 expression, thereby preventing macrophage-mediated phagocytosis and allowing for metastasis. Blocking CD47 enabled macrophage consumption of AML LSCs in vitro, and cleared 90% of AML in the bone marrow of mice transplanted with leukemia. Combining Azacitidine, a drug commonly given to AML patients that increases an ‘eat me’ signal, with a humanized blocking antibody for CD47, led to remarkable results. “Nearly 100% of the patients had some response in MDS or AML… and, as of the time we looked at it, about 50-60% of the patients were now in remission or molecular remission,” Weissman noted. Since different tumor types can express other forms of ‘don’t eat me’ signals, determining the dominant one will be key to well-informed use of this treatment modality.

Targeting CD47 with SIRPαFc Fusion Proteins

Bob Uger develops novel agents targeting the immune system, with a current focus on strategies to inhibit the CD47 ‘don’t eat me’ signal that’s often upregulated by tumor cells. However, “blockade of CD47 alone is not sufficient to drive macrophage phagocytosis,” Uger began. “If we want to make macrophages phagocytic, and eat tumor cells…we also need to deliver a prophagocytic, or ‘eat me,’ signal.” Using the Fc region of an antibody is among the best ways to deliver an ‘eat me’ signal, as it can bind activated Fc γ receptors on macrophages. TTI-621 and TT1-622 are SIRPα-Fc fusion proteins capable of blocking the CD47-SIRPα interaction and delivering an ‘eat me’ signal to macrophages. TTI-621 delivers a strong ‘eat me’ signal with an IgG1 Fc region, while TTI-622 delivers a more moderate signal with an IgG4 Fc region. IgG1 can also activate NK cells.

TTI-621 and TTI-622 block CD47 on tumor cells and stimulate macrophages, which readily present tumor derived antigens to activated T cells.

In an acute myeloid leukemia xenograft model, TTI-621 significantly reduced tumor burden, demonstrating its profound efficacy as a monotherapy. TTI-621 and TTI-622 also synergized well in combinations with small molecule inhibitors and antibody therapies, significantly improving phagocytosis in vitro and impeding tumor growth in vivo. In patients, TTI-621 and TTI-622 exhibited activity across several lymphoma indications, with five and two complete responses observed, respectively. Moving forward, Uger and his team will explore combinations with different classes of agents, as well as efficacy in solid tumors.

BCMA Directed Cell Therapies for Multiple Myeloma

Physician Kristin Hege works to develop innovative therapies to treat challenging cancers. Presently, Hege is focused on targeting B cell maturation protein (BCMA) in multiple myeloma using diverse modalities. One such therapeutic product is IDE-CEL, an autologous CAR T cell transduced with a lentiviral vector encoding a CAR specific for human BCMA. IDE-CEL is the first FDA-approved CAR T cell therapy for multiple myeloma. In relapsed patients, IDE-CEL delivered a high complete response rate. “There was also a trend that suggested an increasing response rate with increasing target dose, with an overall response rate of 82%, and a complete response rate of 39% at the highest dose of 450 million cells,” said Hege. Early and sustained clearance of soluble BCMA (sBCMA), a composite marker of tumor burden, was associated with durability of tumor response.

BB21217 utilizes the same CAR molecule as IDE-CEL, but is cultured with a PI3K inhibitor, resulting in the enrichment of T cells displaying a naïve memory-like phenotype and reduced senescence. BB21217 mediated persistent and durable response lasting up to twenty-four months. Enrichment for memory-like T cells was associated with significantly higher peak CAR T cell expansion and response compared to CAR T cell therapies lacking this feature. Prospective work aims to optimize patient selection for future trials, improve manufacturing processes, and select rational combinations through translational machine learning across CAR T platforms and diseases.

Natural Killer Cells in Solid Cancer

Adelheid Cerwenka researches strategies to exploit NK cells as an adaptive therapy against solid tumors, aiming to address difficulties facing this objective. Among these challenges, the functional impairment of NK cells that have infiltrated solid tumors is perhaps the most notorious. To address this issue, Cerwenka’s team interrogated transcriptional programs of tumor NK cells. A comprehensive analysis uncovered hypoxia inducible factor α (HIF1α) to be highly expressed in tumor NK cells compared to their more functional, blood-based counterparts.

Conditional knockout (cKO) of HIF1α in tumor infiltrated NK cells promoted tumor regression and increased expression of INFγ, Granzyme B, and CD107a. Single-cell sequencing (sc-Seq) of wild-type (WT) and HIF1α cKO NK cells revealed an enrichment of activation molecules and maturation markers in the cKO setting, suggesting that HIF1α negativity enhances the expression of these factors. NFκB family members were also augmented, while checkpoint molecules were downregulated. Further mechanistic exploration demonstrated that IL-18, which was highly enriched in the HIF1α cKO NK cells, played a significant stimulatory role. Neutralization of IL-18 in HIF1α cKO mice precipitated increased tumor growth and decreased NFκB p65, NFκB p50, and IkBζ, indicating that IL-18 mediated improved tumor control and NK cell activation.

Inhibiting HIF1α in human NK cells cultured under hypoxic conditions enhanced effector function. Analysis of sc-Seq data from published databases revealed that positivity of HIF1α in tumor cells negatively correlated with INFγ, demonstrating that loss of HIF1α is essential for INFγ production. These datasets also generated an NK/IL-18/INFγ signature. “The survival probability of patients was much higher when there was a high signature as opposed to a low [signature], indicating the importance of this pathway in different tumor entities,” Cerwenka explained. Hence, HIF1α serves as a checkpoint of NK cell activity, and hypoxia is now noted among the known factors that contribute to tumor NK cell dysfunction and cancer immune escape.

Keynote: Engineering Next Generation Cell Therapies for Enhanced Potency

Physician immunologist Crystal Mackall researches methodologies to improve adaptive cell therapies. Mackall is particularly focused on strategies that increase CAR T cell potency and seeks to address a major question facing the field: is it possible to expand these therapeutics into solid tumors? “I think a logical person looking at the data right now would think it’s not a very promising notion,” Mackall began. “But for those of us who are engaged at the bench [and can see] all of the approaches that are now available to improve the functionality of these cells, we remain optimistic.”

There are significant challenges to expanding immunotherapies for solid tumors.

Mackall’s lab utilized global ATAC-sequencing to assess changes in the epigenome of exhausted human T cells. The most apparent changes identified were in the enhancer landscape, notably regions wherein the transcription factors from the AP-1, Jun, and IRF families bind. Although AP-1, Fos, and Jun are drivers of the functional T cell response, other family members transcribe inhibitory or suppressive gene sets. Overexpressing c-Jun in CAR T cells resulted in significant benefits, including abating hallmark features of exhaustion; preventing terminal differentiation; decreasing expression of checkpoint inhibitors; increasing proliferation; augmenting effector cytokines secretion; and, in models of leukemia and osteosarcoma, promoting tumor regression and long-term survival.

Mackall’s group also explored strategies to “rest” T cells (i.e., transiently cease their signaling) to prevent or reverse the hallmark features of exhaustion. Dasatinib, a potent and reversible inhibitor of CAR-T cell function, downshifted the upregulation of inhibitory signals and transcription factors associated with exhaustion by reprogramming cells at the transcriptional and epigenetic levels. Improved control of tumor burden was also observed in murine models of leukemia and osteosarcoma.

In another resting modality, Mackall’s team implemented a platform wherein a druggable protease, with its cleavage site incorporated into a CAR T cell, would prevent expression at baseline. In the presence of the protease’s inhibitor, the full-length CAR is expressed, and signaling occurs. Compared to constitutively expressed, conventional CAR T cells, SNIP-Trans, a CAR therapy developed on this platform, retained a more balanced CD4+/CD8+ ratio and did not terminally differentiate. The therapy also showed greater functionality when challenged ex vivo in response to plate-bound antigen, showed less expression of CD39, maintained more stem-like and cytotoxic subsets in vivo, and demonstrated enhanced efficacy in several solid tumor models.

“We think that this is a new paradigm for thinking about enhancing immune function in the context of excessive T cell signaling, whether it be due to a tonic signal, or high antigen burden,” Mackall stated.