On December 14, 2020, Arvinas, a pioneer of PROTAC, announced the results of phase I clinical trials of new PROTAC drugs ARV-471 and ARV-110, of which ARV-471 showed outstanding efficacy in patients with ER+/HER2-advanced or metastatic breast cancer who had received various pre-treatments. In the phase 1/2 clinical trial of ARV-110 in the treatment of metastatic castration-resistant prostate cancer (mCRPC), 40% of the patients with specific gene mutations had lower levels of prostate specific antigen (PSA) by more than 50%. The excellent efficacy has attracted the attention of protein researchers.

  1. Mechanism of protein degradation by PROTAC

While theligand binds to POI (protein of interest), E3 ligase-targeting ligand recruits E3 ligase to approach POI to ubiquitinate POI, resulting in POI degradation by endogenous S-protease complex. The X-ray structure of POI-PROTAC-E3 ternary complex also provides strong evidence for this mechanism.

  1. Structure of PROTAC molecule

2.1The ligand of the target protein

In theory, the ligand of the target protein only needs to have a certain affinity with the target protein. The earliest ligands use specific E3 natural substrates, which will lead to excessive molecular weight of PROTAC, single mode of drug administration, poor ability to pass through the cell membrane and so on.

The follow-up ligands are small molecules. Some studies show that some RROTAC molecules formed by ligands with weak binding to POI can still degrade POI with high efficiency. The possible mechanism is that a certain protein-protein interaction between some E3 ligases and POI exists, making POI with weak binding to PROTAC can still form a stable POI-PROTAC-E3 ligase ternary complex for degradation. This discovery cuts down the need for high-affinity ligands when designing PROTAC molecules and expands the scope of potential PROTAC targets.

2.2 E3 recruiting element

There are several types of E3 ligase that can specifically recognize and ubiquitinate proteins by different mechanisms, which are RING/U-box domain family, HECT domain family, and RBR (RING-in-Between-RING) domain family.

E3 ubiquitin ligase: VHL, CRBN, XIAP&clap, Keap1, RNF4, RNF114, MDM2.

2.3 Linker

The design of PROTAC molecules considers retaining a specific site on the two ligands to facilitate the binding to linker. In addition, the characteristics of linker (type, length, attachment position) will affect the stability of E3 ligase.

The results show that the optimum linker length varies from 12 carbon to more than 20 carbon. The commonly used linkers are PEG, alkyl chain, and alkyl/ether chain.

The introduction of hydrophilic PEG can improve the water solubility. If linker is too short, it may lead to steric hindrance, destroy the formation of ternary complex, and reduce the degradation ability of PROTAC. If the linker is too long, the relative motion of both ends of PROTAC will increase and binding constant of E3 and POI will decrease, thus reducing the stability of ternary complex. Other oxygen-containing chains may further change the degradation range of proteins, and oxygen atoms may interact with residues on the ternary binding interface, thus increasing the positive synergistic effect.

In addition, the introduction of heterocycles can reduce the conformational resistance, and the best linker cannot be determined without crystal structure analysis.

2.4 Classification and selection of E3 ligase

More than 600 E3 ligases have been found in the human genome, of which only a few have been used to design PROTAC. The main ligands of CRBN, VHL, MDM2, IAPs, DCAF15, DCAF16, RNF4, and RNF114 E3 ligase have been used in PROTAC development for various cancer targets.

  • PROTACs based on MDM2

The first PROTAC based on MDM2 ligase uses Nutlin-3a as a ligand and non-steroidal selective androgen receptor regulator (SARM), which was developed as an AR ligand to degrade AR.

  • PROTACs based on IAP

The IAP family is a negative regulator of apoptosis, which is composed of eight different members, of which cIAP1, cIAP2, and XIAP are the most known. The inhibition of E3 ligase activity of IAP protein was utilized to develop protein degradation agents. Hashimoto and Naito’s team found the first IAP-based proteolytic agent, which uses methylberstatin (MeBS) as an IAP ligand to target CRABP-1/2. However, in addition to the degradation of their target proteins, these degradants also induce self-ubiquitin and subsequent cIAP1 degradation. To resolve this self-ubiquitination, they replaced ester groups with amide groups at the junction site of MEB, resulting in degradable agents that only degraded CRABP-2 without affecting cIAP1.

  • PROTACs based on VHL

For some targets, PROTACs based on VHL may have better tumor selectivity than PROTACs based on other E3 ligases, such as IAP. For example, the differential expression of VHL between tumor cells and platelets has been demonstrated, which helps to design BCL-XL proteins that retain platelets. This study shows that the binding affinity of VHL ligands does not need to be very high to develop effective PROTACs because the PROTAC with weak binding ligands and VHL proteins and high binding ligands with target proteins can form ternary complexes with target proteins and vhle3 ligases.

On the other hand, VHL-based PROTACs have certain disadvantages. For example, VHL is a tumor suppressor protein that often mutates in some tumors, such as renal clear cell carcinoma or renal cell carcinoma. Therefore, VHL-based PROTACs cannot be used to treat renal cell carcinoma with mutations or deletions of VHL gene. In addition, when using VHL-based PROTACs, the dose needs to be carefully optimized to avoid the tumor suppression function of VHL.

  • PROTACs based on CRBN

Compared with VHL-based PROTACs, CRBN-based PROTACs can reduce a wider range of goals that using the same BCR-ABL ligand, CRBN-based PROTACs degrade BCR-ABL and VHL ABL, while VHL-based ABL can only degrade c-ABL.

Similarly, using the same hybrid kinase inhibitors as warheads, CRBN-based PROTACs degrade more kinases than VHL-based PROTACs. This may be because the protein binding surface of CRBN is much larger than that of VHL, so the significance of CRBN to the formation of ternary complex is greater than that of triggering ubiquitination, which results in more extensive targeting adaptability of CRBN than VHL, and CRBN is widely expressed in tumor cells and normal cells.

Therefore, the possibility of tissue selectivity of CRBN-based PROTACs is relatively small. The advantage of PROTAC based on CRBN is that its molecular weight is relatively lower than that of VHL, so it is more feasible to use CRBN ligands to develop PROTACs with high oral bioavailability.

  • Tissue expression profile analysis of E3 ligase

Histological data make it possible to identify tissue-selective E3 ligases, which can be used to design PROTAC that specifically targets POI in a tissue-selective manner.

Sajid Khan and Yonghan He’s team analyzed more than 50 normal tissues from 1000 individuals, extracted RNA data from tissue expression, and analyzed the expression profiles of E3 ligase genes. The data of 611 E3 ligases were extracted successfully. It was found that some E3 ligases show significant clustering according to their expression levels in some tissues, such as in brain, muscle, and testis. In addition, it was found that 3% of E3 ligases could not be detected, 9% of E3 ligases were low expressed, 4% were highly expressed in almost all normal tissues, about 9% of E3 ligases were specifically expressed in one tissue, and 12% of E3 ligases were enriched in 2-7 tissue types.

These findings suggest that some E3 ligases exhibit tissue selective expression patterns in normal human tissues, which E3 ligases can be used to design tissue-specific or selective pROTAC.