Edward H. van der Horst

Cancer therapies are commonly evaluated by target inhibition, pathway modulation, or initial tumor response, yet these measures do not reliably predict which treatments produce lasting disease control. Many pharmacodynamically active therapies produce transient benefit followed by relapse, while others achieve therapeutic durability despite comparable initial activity. This Perspective argues that durability is better explained by the behavior of the treated tumor system than by the depth of target suppression and distinguishes three modes of intervention behavior: buffered suppression, in which adaptive escape remains broadly accessible; adaptive constraint, in which future tumor states are narrowed without eradication; and collapse-dominant control, in which treatment drives a state transition that leaves few routes for persistence. Examples from targeted therapy, differentiation therapy, checkpoint blockade, and CAR-T therapy illustrate how the same modality can produce different intervention behaviors depending on tumor biology. A prospective sequence is also proposed for testing which behavior a candidate therapy is likely to induce.

PH selective CD28xVISTApH-sens BS2 bsAb inhibits MC38-hVISTA tumor growth in hCD28 KI mice in combination with anti-PD-1.

Full blood cell characterization from ex vivo human whole blood ID.Flow assay.

Tumor Growth Inhibition Summary and Statistics

CD28xVISTA geometries BS2 and BS3 display efficient target binding.

In vitro human T-cell costimulation by CD28xVISTA BS2 potentiates LNCaP killing by a CD3xPSMA T-cell engager.

Expression of VISTA on Kasumi-3 and LNCaP cells.

Reinvigoration of tumor-reactive T cells using costimulatory bispecific antibodies (bsAb) targeting CD28 is emerging as a promising therapeutic strategy. Conditional, tumor-specific recruitment can offer a layer of control and specificity. We developed pH-selective CD28xV-domain Ig-containing suppressor of T-cell activation (VISTA) bsAbs to act specifically within the acidic tumor microenvironment, aiming for enhanced T cell–mediated cancer cell killing while minimizing systemic T-cell activation and cytokine release syndrome risk. CD28 agonism by our CD28xVISTA bsAbs relies on pH-selective engagement of VISTA, a protein robustly expressed on myeloid cells abundant in most solid tumors. Our lead candidate displayed pH-dependent engagement of VISTA and simultaneous binding to CD28, resulting in VISTA-dependent CD28 signaling in a reporter cell line. CD28xVISTA avidly binds VISTA+ cells, and costimulatory activity was shown in vitro by its ability to activate and expand T cells and enhance T cell–mediated cancer cell killing in cocultures of human peripheral blood mononuclear cells and cancer cells in the presence of a tumor-associated antigen–targeted anti-CD3 T-cell engager. This CD28xVISTA bsAb efficiently inhibited the growth of human VISTA-expressing MC38 tumors in a humanized CD28 syngeneic mouse model in combination with PD-1 blockade. Our findings support signaling both in cis (between T cell and target cell displaying peptide–MHC complex) and in trans, with stimulation occurring through CD28 clustering outside of the immune synapse. This CD28xVISTA bsAb showed no signs of superagonistic properties in several in vitro cytokine release syndrome assays. Thus, our data support clinical development for solid tumors in combination with anti–PD-1 or tumor-associated antigen–targeted anti-CD3 T-cell engagers.

Reinvigoration of tumor-reactive T cells using costimulatory bispecific antibodies (bsAb) targeting CD28 is emerging as a promising therapeutic strategy. Conditional, tumor-specific recruitment can offer a layer of control and specificity. We developed pH-selective CD28xV-domain Ig-containing suppressor of T-cell activation (VISTA) bsAbs to act specifically within the acidic tumor microenvironment, aiming for enhanced T cell–mediated cancer cell killing while minimizing systemic T-cell activation and cytokine release syndrome risk. CD28 agonism by our CD28xVISTA bsAbs relies on pH-selective engagement of VISTA, a protein robustly expressed on myeloid cells abundant in most solid tumors. Our lead candidate displayed pH-dependent engagement of VISTA and simultaneous binding to CD28, resulting in VISTA-dependent CD28 signaling in a reporter cell line. CD28xVISTA avidly binds VISTA+ cells, and costimulatory activity was shown in vitro by its ability to activate and expand T cells and enhance T cell–mediated cancer cell killing in cocultures of human peripheral blood mononuclear cells and cancer cells in the presence of a tumor-associated antigen–targeted anti-CD3 T-cell engager. This CD28xVISTA bsAb efficiently inhibited the growth of human VISTA-expressing MC38 tumors in a humanized CD28 syngeneic mouse model in combination with PD-1 blockade. Our findings support signaling both in cis (between T cell and target cell displaying peptide–MHC complex) and in trans, with stimulation occurring through CD28 clustering outside of the immune synapse. This CD28xVISTA bsAb showed no signs of superagonistic properties in several in vitro cytokine release syndrome assays. Thus, our data support clinical development for solid tumors in combination with anti–PD-1 or tumor-associated antigen–targeted anti-CD3 T-cell engagers.

Reinvigoration of tumor-reactive T-cells using co-stimulatory bispecific antibodies (bsAbs) targeting CD28 or CD137 is emerging as a promising therapeutic strategy. Conditional, tumor-specific recruitment can offer a necessary layer of control and specificity. We developed pH-selective CD28xVISTA bsAbs to act specifically within the acidic tumor microenvironment (TME), aiming for enhanced T-cell-mediated cancer cell killing while minimizing systemic T-cell activation and Cytokine Release Syndrome (CRS) risk. CD28 agonism by CD28xVISTA bsAbs relies on pH-selective engagement of VISTA, a protein robustly expressed on myeloid cells highly prevalent in most solid tumors. This modality avoids engagement of tumor-associated antigens (TAAs) with the potential to provide highly tumor specific activity with minimal on-target/off-tumor side effects. We report the identification of a lead candidate with pH-dependent simultaneous engagement of both targets, and VISTA-dependent CD28 signaling in a reporter cell line. CD28xVISTA avidly bound VISTA-positive cells, and co-stimulation was shown in vitro by its ability to activate and expand T-cells and enhance T-cell mediated cancer cell killing in co-cultures of human PBMCs and cancer cells in the presence of a TAA-targeted anti-CD3 T-cell engager. Interestingly, our findings support both signaling in cis (between T-cell and cell displaying peptide-MHC complex) and in trans with stimulation occurring through CD28 clustering outside of the immune synapse. Our lead candidate displayed efficient tumor growth inhibition of human VISTA-expressing MC38 cells in a humanized CD28 syngeneic mouse model in combination with PD-1 blockade. Importantly, our CD28xVISTA bsAb showed no signs of superagonistic properties in several in vitro assays geared towards revealing induction of CRS. Our data supports clinical development in combination with anti-PD-1 or any TAA-targeted anti-CD3 T-cell engagers developed for solid tumors.