Like most drugs, the development interruption of ADCs is typically due to a lack of efficacy, safety issues, commercial considerations, or a combination of these factors. Since 2000, among the 97 ADCs that entered clinical trials and were terminated, the majority (81; 84%) terminated in Phase I or Phase I/II, with only 12 and 4 terminated in Phase II and Phase III, respectively. The majority (67%) contained tubulin-binding payloads, 24% contained DNA-targeting agents (including two calicheamicin payloads and 21 PBD derivatives), and 3% contained topoisomerase 1 inhibitors. Many (80%) targeted tumor antigens that are not representative of approved therapies, while 18 targeted validated targets. Furthermore, 32 ADCs were terminated due to lack of efficacy, 32 due to safety issues, and 29 due to commercial considerations. Several were developed for hematologic malignancies, while the majority were developed for solid tumors.
What lessons can be learned from these terminated projects? Insufficient anti-tumor activity at the maximum tolerated dose seems to be the main reason for termination. Additionally, unacceptable toxicity remains a major obstacle to the development of new ADCs. Better prediction of anticipated adverse events would be another way to reduce premature termination of drug development. For example, the ADC HTK288 targeting cadherin-6 was associated with unexpected central nervous system toxicity, while the ADC LOP628 targeting the tyrosine kinase receptor KIT was associated with unexpected severe hypersensitivity reactions. As demonstrated in the past with cisplatin and paclitaxel, managing the toxicity of a new compound often requires long-term efforts.
Overall, it is challenging to find the right combination of target antigen, active linker-payload, suitable drug-to-antibody ratio (DAR), and appropriate tumor indications. With the rapid increase in validated target antigens and the diversification of payloads, it is expected that further development is needed for some antigen-targeted ADCs that have been terminated.
-
Diversification of payload
Before trastuzumab emtansine was approved in 2019, most ADCs on the market were divided into two types of payloads—agents that bind to tubulin and agents that target DNA. Now, many other drug molecules are being evaluated as potential payloads.
Auristatin derivatives interfere with microtubule polymerization dynamics by disrupting the formation of the mitotic spindle, leading to mitotic arrest and subsequent cell death, thus exerting potent activity. ADCs containing auristatins are currently the largest ADC family.
The second representative class of payloads is DNA-targeting agents, which chemically modify DNA to prevent cell replication. Calicheamicin is a powerful DNA-damaging agent that causes double-strand DNA (dsDNA) breakage through a free radical mechanism and is utilized as the payload in two approved ADCs. PBD dimers are alkylating agents that crosslink dsDNA and are among the most potent, identified cytotoxic agents.
Until recently, most ADCs under development employed potent cytotoxic agents as “warheads”. This was partly due to observations that high DAR values were associated with unfavorable pharmacokinetics, thus favoring the use of more potent payloads that yielded effective ADCs at DAR values ranging from 2 to 4. However, the recently approved T-DXd and trastuzumab deruxtecan with DAR values of approximately 8 suggested the possibility of linking larger amounts of cytotoxic molecules to antibodies without compromising solubility, aggregation propensity, or pharmacokinetic properties. This has led to a paradigm shift, opening the possibility of investigating low-potency compounds with different mechanisms of action as ADC payloads.
Given that not all types of tumors are sensitive to a given type of payload, diversification of payloads is crucial for expanding the indications of ADCs. Over the past two decades, we have witnessed the successful development of ADCs with potent members in these families. Three observations can be made regarding the development of payloads. Firstly, not all commonly used cytotoxic agent families have been successfully utilized as ADC payloads. In particular, attempts to develop nucleoside analogs and antimetabolites as payloads have failed. Secondly, there are currently no approved ADCs that contain payloads with cytotoxic mechanisms distinct from traditional chemotherapy. Finally, among the large number of novel drugs with unique mechanisms of action evaluated in clinical trials, including kinase inhibitors and molecules targeting various processes within tumor cells, many have failed due to poor safety profiles. However, some of these drugs may be potential candidates as ADC payloads. Heidelberg Pharma has brought an example of this type of development into clinical trials, using α-amanitin derivatives as a new ADC payload.
-
Attach importance to toxicity
According to the side effects commonly associated with the types of payloads considered, the toxicity related to ADC administration can be classified as “expected” or “unexpected”.
Expected toxicity: includes peripheral neuropathy induced by MMAE, which is a typical side effect of microtubule-binding agents. Myelosuppression is a common complication of most cytotoxic chemotherapy drugs, especially DNA-targeted drugs.
Unexpected toxicity: includes corneal toxicity associated with the approval of ADCs based on MMAF, such as belantamab mafodotin, even though ADCs based on MMAE are generally not associated with ocular toxicity. Up to 72% of patients exhibit epithelial changes. Additionally, as ADCs are increasingly incorporated into combination regimens, other unforeseen toxicities may be observed. In Hodgkin lymphoma patients, the combination of brentuximab vedotin (BV) with a standard regimen containing bleomycin leads to severe pulmonary toxicity in 44% of patients, while there is no toxicity in the group without bleomycin. The underlying mechanisms for these different toxicities are not fully understood, and they may involve Fc-mediated ADC internalization, release of the payload in normal tissues, off-tumor expression of the target antigen, non-specific uptake of ADCs through phagocytosis or similar cellular processes, enzymatic release of the payload in systemic or specific normal tissue environments, or Fc-mediated inflammatory effects in the context of payload-mediated tissue damage.
As ADCs are increasingly used as first-line or adjuvant therapies, their long-term safety and reversibility of side effects become more and more important. An important issue is the mutagenic effects that certain payloads, especially DNA-targeting agents, may have. Peripheral neuropathy may be irreversible in some patients, highlighting the necessity for individualized dosing and close follow-up in high-risk patients. Another key issue is determining whether certain patients require specific dosing schedules based on variables such as age, sex, types and quantities of prior treatments, comorbidities, or genetic characteristics.
-
Overcoming drug resistance
Considering the series of steps required for successful ADC cytotoxicity, the potential mechanisms of ADC resistance can be complex. ADC resistance can be observed in situations where antigen binding and/or antibody or antigen internalization are reduced, intracellular concentration of the payload is decreased, the target of the payload is altered, or apoptosis mechanisms are modified. Several therapeutic interventions can enhance the efficacy of ADCs in preclinical models. One possibility is to enhance the internalization of ADCs, although the mechanism behind this is not fully understood. Overexpression of caveolin-1, associated with lipid rafts, has been shown to increase the internalization of T-DM1. In cases where resistance has already developed, a second type of payload can be used. Antigen expression heterogeneity is a classical mechanism of antibody-based treatment failure. Preclinical models have demonstrated that antibody distribution in tumors depends on the expression of the target antigen. The bystander effect provided by most ADCs, such as T-DXd, involves the release of unbound payloads in the tumor microenvironment and their potential uptake by adjacent tumor cells independent of their antigen profile. This is a potent strategy against tumor antigen heterogeneity, as the payload can reach tumor cells with low levels of the target antigen. Furthermore, the administration route of ADCs is expected to strongly influence the occurrence of resistance. In early trials, ADCs were primarily administered as monotherapy, which can promote the selection of resistant tumor populations. Currently, many combination regimens are being explored in clinical practice, including traditional cytotoxic chemotherapy and other targeted agents. The treatment sequence may also be an important parameter.
-
Combination therapy
Brentuximab vedotin (BV) has been combined with over 80 different regimens, including cytotoxic chemotherapy and immune checkpoint inhibitors. The combination of BV with immune checkpoint inhibitors shows great promise. After it was proven that single-agent pembrolizumab was effective in patients with relapsed Hodgkin lymphoma and that it was more effective than BV in this situation, some studies looked into combining BV with anti-PD-1/PD-L1 or anti-CTLA-4 drugs. In patients with relapsed or refractory Hodgkin lymphoma, the combination of BV with nivolumab induces an 82% ORR, including a 61% complete response (CR).
T-DM1 has also been investigated in combination regimens. The addition of capecitabine to T-DM1 does not improve ORR but leads to more adverse events. T-DM1 in combination with pertuzumab yields contrasting results. In the MARIANNE study, patients with advanced breast cancer receiving T-DM1 in combination with pertuzumab had similar overall survival but better quality of life compared to patients receiving trastuzumab and paclitaxel. T-DM1 has also been used in combination with immune checkpoint inhibitors, with support from clinical and preclinical data.
Overall, these studies suggest that carefully selected combinations of ADCs with other drugs may be superior to non-conjugated antibody-based therapies in terms of patient outcomes and safety. Safety is a key concern in combination design, particularly for frail or comorbid patients.