The structure of antibody-drug conjugates (ADCs) is composed of three basic components, namely, tumor-specific monoclonal antibodies (mAb), linkers, and cytotoxic drugs (toxin) which we usually call payloads or warheads. ADC drugs use the fine specificity of tumor-specific mAbs to deliver highly viable cytotoxic drugs to the tumor site, thereby exerting a tumor-killing effect. The tumor-killing effect and side effects of ADC drugs largely depend on the payload. An ideal payload drug needs to have the following four points: high cytotoxicity, small molecular weight, clear mechanism of action, and can be modified. There is currently no universal payload. Researchers in the field mainly consider the following points when developing payload: tumor type and microenvironment, expression level of target protein, ADC molecular design (coupling site, DAR value, etc.).

What is Tumor Heterogeneity?

Tumor heterogeneity is one of the characteristics of malignant tumors. It refers to that during the growth of tumors, after multiple divisions and proliferations, their daughter cells show molecular biological or genetic changes, so that the growth rate, invasion ability, drug sensitivity, prognosis and other aspects of tumors are different. Simply put, tumor heterogeneity refers to the presence of both tumorigenic cell subpopulations and non-tumorigenic cell subpopulations within a tumor. The heterogeneity of solid tumors is more obvious, which is also a difficulty in the treatment of solid tumors.

What is Bystander Effect?

The bystander effect means that some ADCs also have anti-tumor activity against other tumor cells surrounding the target target antigen tumor cells, regardless of whether these cells express the target antigen. The bystander effect may work in two ways:

• Release the payload before ADC internalization;
• Intracellular payloads can diffuse across the cell membrane.

This pharmacological profile is closely related to the structural design of the ADC, especially the biochemical features of the linker and payload. The chemistry of the linker that connects the drug to the mAb determines when and how the drug comes off the antibody. The different chemical properties of the payload determine whether different payload molecules can diffuse into the surrounding tumor tissue and thereby exert a cell-killing effect. Currently, there are also preclinical studies that have designed mathematical models to evaluate the extent of the bystander effect.

Mechanism of Bystander Killing Effect

According to research, the bystander killing effect mainly includes the following possible mechanisms:

• ADC enters the cell through endocytosis after binding to the target antigen. Under the action of the corresponding enzymes in the lysosome, the ADC is degraded and the toxin molecules are shed. Then they penetrate the cell membrane and are released, and enter neighboring cells to exert a killing effect;
• After ADC binds to the target antigen, it undergoes endocytosis before being cleaved by enzymes such as cathepsin-B outside the tumor cells; toxin molecules are directly released extracellularly and penetrate the membrane into adjacent cells to exert a killing effect;
• In addition, the target tumor cells bound by ADC may be internalized through Fc-mediated phagocytosis, and the toxin molecules penetrate the cell membrane and are released, and penetrate the membrane into neighboring cells to exert a killing effect.

Fig. 1. Mechanism of bystander killing effect of ADC drugs (Br J Cancer. 2017, 117(12): 1736-1742).

Payload with Bystander Effect

The bystander killing effect requires that the linker must be cleavable to ensure the release of toxin molecules, but it is not necessarily necessary for endocytosis. Potential payloads with bystander killing effects generally require good membrane permeability. Generally, such molecules are required to be neutral, uncharged molecules with strong hydrophobicity. However, hydrophobicity also needs to be balanced. If the hydrophobicity is too strong, ADC drugs will easily aggregate or be absorbed by tissues. They will be too toxic and not suitable for the development of ADC drugs. Based on relevant information from public channels, it was found that payloads with bystander killing effects mainly include: MMAE, DXD, and Eribulin; payloads without bystander killing effects mainly include: Amanitin, MMAF (negatively charged).

Tubulin Inhibitors

Tubulin inhibitor payloads mainly act on cells during mitosis, changing the cytoskeletal structure of cells, causing tumor cell death, and have no killing effect on non-dividing tumor cells. The most classic payloads are maytansine derivatives represented by DM1 and DM4 and auristatin derivatives represented by MMAE and MMAF. This kind of payload is relatively mature, but it mainly has disadvantages such as a long half-life, a narrow anti-tumor spectrum, and a large number of targets. In contrast, MMAE/MMAF is more widely used because its cytotoxicity is stronger than that of DM1. Compared with MMAE, MMAF can produce metabolites with negatively charged carboxyl-terminal phenylalanine residues, which are unable to penetrate the membrane and have limited applications. Therefore, MMAE is more widely used among the two.

DNA Inhibitor

DNA inhibitors can act on the entire cell cycle and have therapeutic effects on solid tumors. Because the amount of DNA isomerase in tumor cells is small, only a smaller dose of payload is needed to achieve the killing effect compared to tubulin inhibitors. DNA inhibitors mainly include antromycin derivatives (PBD), calicheamicins, duocarmycin (SYD985), and camptothecin topoisomerase I inhibitors (Dxd, SN-38). DNA inhibitors generally have a short half-life and are not prone to accumulation in the body. For example, the half-life of SN-38 is only 6.36 h, and the half-life of Dxd is even shorter, only 1.37 h.

ADC Products from BOC Sciences

ADCs with Bystander Effect

DS-8201

As a new generation of ADC drugs, DS-8201a's unique structural and mechanism properties endow it with powerful anti-tumor activity. A cleavable GGFG tetrapeptide linker was used to couple the permeable highly active toxin molecule Dxd with trastuzumab through cysteine, and a strong bystander killing effect was obtained. In preclinical studies (shown in Fig. 2), each HER2-targeting ADC or control ADC was administered to co-vaccinated xenograft mice, and tumor volume and luciferase activity were measured for each mouse. In the DS-8201a group, a significant decrease in luciferase signal was observed, indicating that MDA-MB-468-Luc cells were completely eliminated by DS-8201a. Similar results were not observed with other HER2-targeting ADCs or control ADCs. In terms of tumor volume changes, anti-HER2-DXd (2) and T-DM1 inhibited tumor growth, but not as effectively as DS-8201a (Fig. 2c), which may be attributed to the limited clearance of tumors, especially by anti-HER2-DXd alone (2) and T-DM1 detection in HER2-positive populations. In the anti-HER2-DXd(2) treatment group and T-DM1 treatment group, immunohistochemical analysis showed that HER2-positive cells were eliminated and HER2-negative tumor cells occupied the majority of the tumor tissue. In the DS-8201a treatment group, nearly all HER2-positive and HER-negative tumor cells disappeared, and the tumors were almost free of tumor cells. These results confirmed that DS-8201a showed anti-tumor activity not only against HER2-positive tumor cells but also against HER2-negative tumor cells under combined vaccination conditions, while T-DM1 hardly killed HER2-negative tumor cells.

Fig. 2. Bystander killing effect under in vivo coinoculation conditions (Cancer Sci. 2016, 107(7): 1039-46).

REGN3124-PBD

REGN3124 is conjugated to a PBD dimer via a cathepsin B-cleavable dipeptide (Val-Ala) linker and is used in preclinical studies to treat glioblastoma. Studies have shown that aberrant EGFRvIII signaling plays an important role in tumor development and that previous treatments failed in part from the development of EGFRvIII-negative tumor cells in heterogeneous tumors. REGN3124-PBD developed by Regeneron uses the bystander killing effect to target glioblastoma (GBM) that heterogeneously expresses EGFRvIII. The cytotoxicity and bystander killing effects of drugs were detected in subcutaneous xenograft models of U251/EGFRvIII and U87/EGFRvIII cell lines (both U87 and U251 are GBM cell lines), as well as subcutaneous xenograft models of EGFRvIII-positive patients (GBM6 and GBM59). The results showed that a single dose of 0.38 mg/kg REGN3124-PBD could induce the continuous regression of subcutaneous U251/EGFRvIII and U87/EGFRvIII xenografts, and a single dose of 0.53 mg/kg REGN3124-PBD could make the subcutaneous patient-derived xenograft GBM6 tumor completely regress and GBM59 tumor continue to regress.

As the competitive landscape of ADC drugs further develops, the development of new payloads will play an increasingly important role. Given the importance of payloads with bystander killing effects on heterogeneous tumor cells, the development of such payloads is a good direction. However, it is undeniable that the bystander killing effect may also lead to non-specific killing of normal cells. Therefore, the payload and linker need to be reasonably selected and designed based on the actual situation of the target and specific indications to reduce or avoid the adverse effects of the bystander killing effect and give full play to the bystander killing effect.