1. Selection of the correct target
The successful development of ADC relies on the specific binding of antibodies to the target antigen. The ideal ADC target is highly expressed on the surface of tumor cells, low or not expressed in normal tissues, or at least limited to specific tissues, such as CD138, 5T4, mesothelin, leukemia and CD37. Targets expressed in normal tissues will be taken up by ADC drugs, leading not only to “off-target” toxic effects, but also to lowering the dose of ADCs enriched in cancer tissues and reducing the therapeutic window of ADC drugs.
Effective ADC activity is related to the number of antigens on the cell surface. Studies have shown that to achieve effective ADC activity, at least 104 antigens on the cell surface are required to ensure that a lethal dose of cytotoxic drugs is delivered to the interior of the cell. Due to the limited number of antigens on the tumor cell surface (approximately 5,000 to 106 antigens per cell surface on average), and the average DAR of most clinical stage ADC drugs is 3.5 to 4, the delivery of ADC drugs into tumor cells is limited. rare. This is also considered to be one of the main reasons for the clinical failure of ADCs combined with conventional cytotoxic drugs such as methotrexate, paclitaxel and anthracyclines.
In addition to specificity and adequate expression, the optimal target antigen should also lead to efficient ADC internalization. Binding of antibodies to target cell surface antigens can trigger the internalization pathway of antibody-antigen complexes into cells, enabling intracellular delivery of drugs.
At present, leukocyte surface differentiation antigen is the first widely used ADC target, and 10 of the 20 ADC drugs currently in clinical development (CD33, CD30, CD79b, CD22, CD19, CD56, CD138, CD74) are leukocyte surface antigen. Many ADC drugs target leukocyte surface antigens in large part because these antigens are highly expressed in tumor tissues and not expressed in normal hematopoietic tissues, or expressed at extremely low levels.
In addition, some solid tumor surface receptor molecules have gradually been found to be suitable clinical ADC targets, such as PSMA on prostate cancer, epidermal growth factor receptor EGFR and ovarian cancer tissue nectin 4 and other ADC drugs have entered the clinic Phase II. Kadcyla, which was approved by the FDA in 2013, targets HER2. Padcev, which was approved by the FDA in 2019, targets NECTIN4, which is the second ADC drug target approved for the treatment of solid tumors.
2. Selection of Antibodies
The high specificity of antibody molecules is an essential requirement to achieve the efficacy of ADC drugs, thereby concentrating cytotoxic agents at the tumor site. Relying on high-affinity specific antibodies, in addition to avoiding toxicity to healthy cells, antibodies lacking tumor specificity may be eliminated by the circulating system, resulting in “depletion” of ADC drugs before they reach the tumor tissue. To this end, cytotoxic drugs are typically attached to the Fc portion or constant region of the mAb to prevent any effect on antigen detection and binding.
Since these 150kDa antibody molecules not only contain multiple natural sites for conjugation, but can also be modified for other reactive sites, all ADC antibodies currently are IgG molecules. The advantages of IgG molecules are their high affinity for the target antigen and longer half-life in the blood, which leads to their increased accumulation at tumor sites. Compared with other IgG molecules, IgG1 and IgG3 have much stronger antibody-dependent cytotoxicity (ADCC) and complement-dependent cytotoxicity (CDC), but IgG3 is not ideal for ADC drugs due to its short half-life. In addition, compared with IgG2 and IgG4, the intracellular hinge formed by IgG1 is easy to reduce, so it is difficult to produce ADC drugs based on cysteine production. Therefore, since IgG1 has relatively strong ADCC and CDC, long half-life, and easy production, most ADC drugs are currently constructed using IgG1 scaffolds.
The immunogenicity of ADCs is one of the major determinants of circulating half-life. Early ADCs used murine monoclonal antibodies to induce a strong, acute immune response (HAMA) in humans, and most ADCs currently use humanized or fully humanized antibodies.
Overall, an ideal mAb for an ADC architecture should be a humanized or fully humanized IgG1 molecule capable of selectively binding tumor cells without cross-reacting with healthy cells. Furthermore, ADC internalization may be an important rather than an absolute factor for successful treatment.
3. Selection of toxin molecule (Payload)
Toxin molecules are the key factor for the success of ADC drug development. Only a small part of the injected antibody accumulates in solid tumor tissue, so it is necessary to have sub-nanomolar toxic molecules (IC50 value of 0.01-0.1nM) first. Appropriate payloads. In addition, the toxic molecule must have suitable functional groups that can be conjugated, be strong cytotoxic, be hydrophobic, and be very stable under physiological conditions.
The toxic molecules currently used for ADC drug development can be divided into two categories: microtubule inhibitors and DNA damaging agents. Other small molecules, such as α-amanitin (selective RNA polymerase II inhibitor), are also under investigation. The former is represented by MMAE and MMAF (free drug IC50: 10-11-10-9M) of Seattle’s Genetics and DM1 and DM4 (free drug IC50: 10-11-10-9M) developed by ImmunoGen’s company. The latter is represented by the PBD (free drug IC50<10-9M) of Calichemicin, duocarmycins, and Spirogen’s. These toxins have corresponding ADC drugs for exploration and development in the clinical stage. Many companies are also developing their own payloads, such as Nerviano Medical sciences, MersanaTherapeutics and other companies.
4. Selection of Linker
Although it is important to select specific antibodies and payloads according to the type of tumor cells, constraining antibodies and payloads by choosing the appropriate linker is the key to successfully constructing ADCs in terms of pharmacokinetics, pharmacology and therapeutic window, ideal linker The following conditions must be met:
(1) The linker needs to exist stably in the blood circulation system, and can rapidly release active payloads when positioned in or near tumor cells. The instability of the linker will lead to the premature release of payloads, causing damage to normal tissue cells. There is also a clinical study showing an inverse relationship between the ADC stability of bismuth alkaloids and adverse reactions. Therefore, it is important to identify the linker with the best stability for the combination of antibody, tumor tissue and payload.
(2) Once the ADC is internalized into the target tumor tissue, the linker needs to have the ability to be rapidly cleaved and release toxic molecules.
(3) Hydrophobicity is also an important feature to consider in linkers. Hydrophobic linking groups and hydrophobic payloads usually promote the aggregation of ADC small molecules, thereby causing immunogenicity.
At present, linkers are divided into two categories according to whether they are: one is cleavable linker (acid-labile linkers, protease cleavable linkers, disulfide linkers), the main type of ADC drug; the other is uncleavable linker, the difference is whether it will degraded in cells.
The designed cleavable linker takes advantage of its environmental differences in the blood system and tumor cells. For example, acid-sensitive linkers are usually very stable in blood, but unstable in lysosomes at low pH, and degrade rapidly, releasing free Active toxic molecule (Mylotarg (gemtuzumab ozogamicin)). Likewise, protease-sensitive protease cleavable linkers are stable in blood, but are rapidly cleaved in lysosomes rich in proteases (recognizing their specific protein sequences) to release active toxic molecules, as in Val-Cit dipeptide cross-linkers. The linkage is rapidly hydrolyzed by intracellular cathepsins enzymes (Adcetris (brentuximab vedotin)). The designed disulfide-crosslinked linker takes advantage of the high-level expression of intracellular reduced glutathione, which releases a toxic molecule (IMGN-901 (anti-CD56-maytansine)) intracellularly.
The non-cleavable linker is composed of stable bonds that are resistant to protease degradation and is very stable in blood. It relies on the complete degradation of the ADC antibody component by cytoplasmic and lysosomal proteases, and finally releases the payload linked to amino acid residues derived from the degrading antibody for killing. Cancer cells (eg ado-trastuzumab emtansine, T-DM1, or Kadcyla). At the same time, ADC drugs that cannot be cleaved by linker cannot be released extracellularly, and cannot kill nearby cancer cells by “bystander effect”.
Of course, which type of linker to choose is closely related to target selection. Among ADC drugs with cleavable linkers, those targeting B cell antigens (CD19, CD20, CD21, CD22, CD79B, CD180) have been shown to be very effective in vivo. In contrast, among ADC drugs with non-cleavable linkers, targets confirmed to be endocytosed in vivo and rapidly transported to lysosomes include CD22 and CD79b.
Ensuring the specific release of free drugs in tumor cells is the ultimate goal of choosing Linker, and the control of drug toxicity is also very important. Ultimately, a case by case analysis is needed to determine how to optimize the selection of appropriate linkers, targets and toxic molecules to balance the effectiveness and toxicity of ADC drugs.
