Targeting of the tumor necrosis factor receptor superfamily for cancer immunotherapy

Abstract

The tumor necrosis factor (TNF) ligand and cognate TNF receptor superfamilies constitute an important regulatory axis that is pivotal for immune homeostasis and correct execution of immune responses. TNF ligands and receptors are involved in diverse biological processes ranging from the selective induction of cell death in potentially dangerous and superfluous cells to providing costimulatory signals that help mount an effective immune response. This diverse and important regulatory role in immunity has sparked great interest in the development of TNFL/TNFR-targeted cancer immunotherapeutics. In this review, I will discuss the biology of the most prominent proapoptotic and co-stimulatory TNF ligands and review their current status in cancer immunotherapy.

Figure 1

TNFL/TNFR signaling characteristics. (a) TNF-ligands are typically produced as type II transmembrane proteins, but the extracellular domain of most of these ligands can also be proteolytically cleaved by proteases, such as ADAM-17 (also known as TACE) [10], into a soluble form. Typically, the soluble ligand retains binding activity but has lost some or all receptor-activating activity. This activity can be restored by secondary cross-linking. (b) Signaling requirements of 4-1BB-signaling by s4-1BBL. (c) The cross-linking requirement of sTNF ligands makes their inclusion into an antibody fragment approach attractive. In brief, such a TNFL-fusion protein comprises a scFv antibody fragment genetically fused to the TNFL. This scFv:TNFL fusion protein is essentially inactive en route. However, upon target binding of the scFv antibody fragment domain the soluble ligand is converted into a signaling competent membrane-like ligand. (d). Illustration of target cell-restricted activation by scFv:4-1BBL.

Figure 2

TNF/TNFR signaling and TNFR-targeted therapeutics. (a) TNFR1 and TNFR2 are effectively activated by membrane TNF, but sTNF can only trigger TNFR1-signaling. (b) TNFR-targeted drugs include a stabilized TNFR2-selective scTNF that may help to induce TNFR1 proapoptotic signaling, as well as targeted strategies such as scFv:sTNFL, and scFv:sTNF-TNFR1 prodrug constructs. The latter only become activated after target antigen-selective binding and subsequent cleavage of the TNFR1 inhibitory domain by tumor-overexpressed proteases.

Figure 3

Design of FasL/Fas-based cancer therapeutics. (a) Soluble homotrimeric FasL is essentially incapable of activating Fas-apoptotic signaling. However, hexamerized recombinant forms of sFasL have Fas-activating capacity analogous to membrane-expressed FasL. (b) The inactivity of homotrimeric sFasL has been exploited in scFv:FasL fusion proteins, by which the full apoptotic potential of FasL/Fas signaling is unleashed only upon target antigen-selective binding. To further increase the safety of FasL-based therapeutics, a FasL-based prodrug strategy analogous to TNF has been designed and evaluated.

Figure 4

TRAIL/TRAIL-receptor signaling and design of TRAILR agonists. (a) Membrane expressed TRAIL triggers apoptotic signaling via TRAIL-R1 and TRAIL-R2, whereas soluble TRAIL only efficiently activates TRAIL-R1. Recombinant nontargeted sTRAIL thus predominantly triggers TRAIL-R1 apoptotic signaling. Recombinant TRAIL-R1 or TRAIL-R2 agonistic antibodies can selectively activate TRAIL-R1 or TRAIL-R2, respectively. (b) Tumor-targeted delivery of sTRAIL, using scFv:sTRAIL, results in conversion of sTRAIL to membrane-like TRAIL that can induce apoptosis via TRAIL-R1 and TRAIL-R2. The antibody fragment may inhibit or activate target antigen signaling and thereby contribute to the antitumor activity of scFv:sTRAIL. (c) Targeting of T-cell markers CD7 or CD3 with K12:TRAIL and antiCD3:TRAIL, respectively, equips T-cells with membrane-like proapoptotic TRAIL that enhances antitumor T-cell activity. The antiCD3 scFv can also trigger stimulatory signaling in resting T-cells and trigger granzyme/perforin-mediated cytotoxicity.

Figure 5

Co-stimulatory TNF ligands provide crucial signals for generation of antitumor T-cell responses. Within the antitumor T-cell immune response, CD40L-mediated CD40 costimulation of DCs by CD4+ T-helper cells is critical for the generation of CD8+ T-cell responses. 4-1BBL expressed on DCs stimulates the generation of T-cell response, while at the same time inhibiting the formation of inducible regulatory T-cells in the tumor micro-environment. CD70 serves to efficiently prime CD4+ and CD8+ T cell responses, to enhance T-cell survival, and to optimize effector function. OX40 is transiently upregulated upon T-cell activation and enhances clonal expansion, survival, proinflammatory cytokine production, and generation of memory CD4 T-cells and enhances CD8+ T-cell survival and expansion.

Figure 6

CD40L/CD40-based agonists for cancer therapy. Soluble CD40L is only capable of sub-optimal CD40 signaling. However, enforced trimerization of CD40L or agonistic antiCD40 antibodies can trigger effective CD40-signaling, but with severe side-effects due to ubiquitous CD40-activation. Of note, CD40 agonist antibodies require FcR-mediated cross-linking for effective CD40-signaling. In an antibody fragment-targeted scFv:CD40L fusion protein, the CD40L domain is relatively inactive en route, but gains membrane-like activity upon target antigen-mediated binding. Further, CD40L can be virally transduced into tumor cells, using AdCD40L, to optimize CD40-mediated co-stimulation.

Figure 7

CD70/CD27-based agonists for cancer therapy. CD70 is highly expressed on malignant cells and thus a bona-fide target for antiCD70 antibody based therapy. Similarly, scFv-targeted TRAIL/FasL-based fusion proteins could be used to selectively deliver and locally activate proapoptotic signaling. Triggering CD27 T-cell co-stimulatory signaling may be particularly applicable in, for example, ex vivo expansion of adoptive T-cells. Incorporation of an antiCD27 scFv in a bispecific antibody format with a tumor-specific targeting antibody fragment may open up ways to ensure selective modulation and/or inhibition of CD27 signaling in the tumor micro-environment.

Figure 8

4-1BBL/4-1BB. Soluble 4-1BBL or hexameric 4-1BBL is essentially inactive. However, enforced oligomerization using SA-4-1BBL enables 4-1BB activation with a favorable toxicity profile. In contrast, agonistic 4-1BB antibodies potently activate 4-1BB signaling but with dose-limiting toxicity. The selective use of 4-1BB co-stimulatory signaling can potentiate CAR T-cell activity and trigger effective lysis. Further, selective delivery of 4-1BBL using scFv:4-1BBL ensures target antigen-restricted conversion of inactive s4-1BBL into membrane-like and signaling competent 4-1BBL.

Figure 9

OX40L/OX40. Soluble OX40L can only suboptimally activate OX40 co-stimulatory signaling. Hexameric recombinant OX40L is fully capable of activating OX40-signaling like OX40 agonistic antibodies, with no dose-limiting toxicity of such an OX40 antibody in an early clinical trial. To increase tumor selectivity, sOX40L can be targeted to tumor cells using scFv:OX40L. The sOX40L domain will convert into membrane-like and fully signaling competent OX40L only upon selective binding to the targeted antigen.