Allogeneic hematopoietic stem cell transplantation (HSCT) is the original cellular therapy. When HSCT is used for treatment of patients with leukemia and other hematologic malignancies, clinicians hope that polyclonal T cells from healthy donors will recognize tumor cells that the patient's own T cells do not, thereby exerting an anti-neoplastic "graft-versus-leukemia" (GvL) effect. For decades, oncologists have also used high-dose Interleukin-2 to non-specifically activate large numbers of T cells in situ and achieve long-term remissions in patients with melanoma and renal cell cancer (RCC). However, these approaches often fail.
From these beginnings, the field of immuno-oncology has recently exploded with novel therapies that more elegantly exploit the tumor-antigen recognition and cytotoxic potential of T cells. In broad strokes, these approaches 1) bring T cells and cancer cell targets into physical proximity, 2) increase numbers of T cells capable of recognizing a particular tumor antigen, 3) modulate T-cell activity once relevant tumor-associated antigen is encountered, 4) make tumors more attractive or accessible targets for cytotoxic T cells, and 5) disable the "brakes" or inhibitory signals that limit magnitude or duration of activity against neoplastic cells following T-cell activation (Figure 1).
1. Blinatumomab, a two-headed, bi-specific T-cell engager (BiTE) antibody physically links a CD19+ B cell with a CD3+ T cell to facilitate potential recognition of CD19+ neoplastic cells by cytotoxic T cells, with clinical response rates of 30-80% in relapsed B-cell acute lymphoblastic leukemia (ALL). However, proximity does not guarantee antigen-specificity of the snared T cell, duration of response to blinatumomab is limited, neurotoxicity and cytokine release can be problematic, and relapses with CD19 negative sub-clones can occur.
2. Increased specificity of T-cell responses can be accomplished through genetic engineering in which an antigen-specific receptor fused to an activating domain (e.g., a chimeric antigen receptor, CAR) is introduced and results in CAR protein expression on a large number of T cells. As exemplified by the success of early phase trials of CD19-directed CAR T cells in B-cell non-Hodgkin lymphoma (B-NHL), chronic lymphocytic leukemia (CLL), and ALL, activation of CAR T cells can lead to massive T-cell expansion, cytokine release, and death of malignant cells, with complete response rates of up to 90 percent in some settings.
3. "Second-generation" CAR T cells in current clinical trials are characterized by inclusion of a costimulatory domain to facilitate transmission between antigen-recognition and CD3ζ activation domains, which has proven necessary for significant clinical efficacy. Future generations of CAR cells, TRUCKs (T cells Redirected for Universal Cytokine-Mediated Killing), and TANKs (Target-activated Natural Killers), which are now in development, are being engineered to express additional costimulatory domains to enhance signaling; deliver cytokine "payloads" to assist in activation and homing; resist inhibition by soluble or cell-surface proteins in the tumor microenvironment; encode "on/off" switches that can be triggered by small molecules administered at the bedside to temporally modulate T-cell activity and ameliorate toxicity; and even leverage natural killer cell (NK) cytotoxicity (Figure 2).
4. To enhance presentation of tumor antigens by professional antigen-presenting cells and augment activity of cytotoxic T cells in situ, numerous cancer vaccines are in development, particularly for patients with solid tumors. These vaccines may consist of autologous dendritic cells (DCs) loaded in vitro with proteins obtained from a tumor cell line; autologous tumor cells (e.g., obtained during primary ovarian cancer debulking) transduced ex vivo with a viral vector encoding GM-CSF to facilitate in vivo DC homing and maturation; or autologous tumor proteins compounded with CpG Toll-like receptor activators and GM-CSF in a biodegradable polymer matrix to create an optimized 3-dimensional environment for tumor antigen, DC, and T-cell interactions.
5. Immune checkpoint inhibitors, such as those targeting CTLA-4 or PD-1/PD-L1 interactions are demonstrating striking clinical success by preventing down-regulation of T cells that have already been activated and prolonging the lymphocyte attack against tumor cells (see articles by Dr. Armand, Dr. Davids, and Drs. Bianchi and Munshi in this issue of the newsletter).
These approaches are only the beginning of the possible uses of immunotherapy in cancer treatment. CARs relying on antibodies for recognition of tumor cells are being joined by chimeric constructs utilizing activating receptors found on NK cells (Figure 3). While CARs recognize extracellular antigens on a tumor surface, T cells engineered to express additional T-cell Receptor (TCR) proteins can recognize a totally different pool of intracellular tumor antigens. Engineered TCR cells have the potential to reduce "on target antigen" but "off tumor cell" effects such as aplasia of normal B cells are frequently seen after CD19 CAR therapy. Additionally, complex, time-intensive genetic engineering is not always necessary to boost antigen-specific responses. In vitro expansion of pre-existing/natural antigen-specific T cells can be sufficient. Demonstrated efficacy of Epstein-Barr virus (EBV) or cytomegalovirus (CMV)-specific cytotoxic T lymphocyte (CTL) clones in post-HSCT infections or viral-driven lymphomas has spawned trials of CTLs directed against tumor-antigens such as PRAME and NY-ESO 1 as therapies for relapsed/refractory NHL.
These immunologic modalities can be combined. For example, CAR T-cell infusion can be followed by a checkpoint inhibitor, potentially bringing to bear a large number of antigen-specific and easily-activated T cells while concurrently disabling inhibitory signals that tumors have relied on for immune evasion. These creative manipulations of both adoptively transferred and native immune cells can be applied to a broad spectrum of diseases. The clinical trials referred to in Figure 1 are only the beginning. We at Dana-Farber are excited to be part of the ever-expanding 360-degree frontier of cellular therapy and immuno-oncology through our clinical care, novel therapeutic trials, and on-site manufacturing of cell therapies in the Dana-Farber Cell Manipulation Core Facility.
Figure 1. The 360 Degrees of Immuno-oncology
Surrounding the central tumor cell are schematics representing approaches to optimize T cell activity and recognition in a non-antigen-directed way (in blue), by adoptively transferring antigen-specific T cells such as CAR T cells (in green), or by optimizing activation and expansion of tumor-antigen-recognizing T cells already present in the patient via cellular vaccines (in red). Multi-colored circles – polyclonal T cells. Yellow haloes – polyclonal T cell activation. Triangles – cytokines. Purple circles – T cells specific for tumor antigens. Purple haloes – antigen-specific T cell activation. In sector 4 – irradiated (red X's) AML blasts transfected with adenovirus (blue rings) encoding GM-CSF (triangles) and dendritic cells loaded with tumor antigen (purple dots). Numbers reflect the numbered approaches in the text. * indicates diseases targeted in active or recent trials conducted at Dana-Farber. AML – acute myeloid leukemia; HD – Hodgkin's Disease; NSCLCa – non-small cell lung cancer; GU – genitourinary cancers; HNSCC – head and neck squamous cell cancer; GBM – glioblastoma multiforme; FL – follicular lymphoma; CRC – colorectal carcinoma.
Figure 2. Ongoing CAR Engineering
Current 2nd generation CAR constructs contain an scFv antibody-derived recognition domain (2 purple ovals, comprised of immunoglobulin light and heavy chain regions), with a variable-length hinge region (black solid line), single costimulatory domain (dark blue), and CD3ζ activation domain (green). Light blue rectangles – additional costimulatory domains. Cytokines – yellow triangles. Grey box – "logic" introduced by interaction between 2 separate or "split" receptors recognizing different antigens. Black semicircles – binding moieties to allow separately encoded antigen-recognition/costimulation and activation domains to effect T-cell activation only in the presence of a small molecule (yellow diamond). Black dotted lines – T cell membrane. Red dotted line – NK cell membrane.
Figure 3. NK Receptor and TCR-based Engineering
Antigen recognition need not be restricted to scFv domains but can be carried out by NK-cell receptors (top tight panel, red) or T-cell receptors (TCRs) (bottom panel). These may be engineered and introduced in a construct containing costimulatory and activating domains or may colocalize with proteins serving those functions naturally expressed by the T cell. TCRs can recognize intracellular tumor antigens presented in the context of major histocompatibility complexes (MHCs), although this restricts utility to patients of a specific MHC haplotype. Orange, light purple and light blue rectangles, bottom left panel – natively expressed TCRs. Dark purple rectangles, bottom middle panel – engineered TCRs introduced into each T cell in addition to native TCR. Light purple rectangles, bottom right panel – EBV-specific TCRs, for example, that are present naturally but can be expanded against viral peptides in vitro and reinfused. (CD3δ, ε, γ and CD4 and CD8 molecules are not included in schematics for simplicity.)