Novel approaches to targeted protein degradation technologies in drug discovery

Abstract

Introduction:

Target protein degradation (TPD) provides a novel therapeutic modality, other than inhibition, through the direct depletion of target proteins. Two primary human protein homeostasis mechanisms are exploited: the ubiquitin-proteasome system (UPS) and the lysosomal system. TPD technologies based on these two systems are progressing at an impressive pace.

Areas Covered:

This review focuses on the TPD strategies based on UPS and lysosomal system, mainly classified into three types: Molecular Glue (MG), PROteolysis Targeting Chimera (PROTAC), and lysosome-mediated TPD. Starting with a brief background introduction of each strategy, exciting examples and perspectives on these novel approaches are provided.

Expert Opinion:

MGs and PROTACs are two major UPS-based TPD strategies that have been extensively investigated in the past decade. Despite some clinical trials, several critical issues remain, among which is emphasized by the limitation of targets. Recently developed lysosomal system-based approaches provide alternative solutions for TPD beyond UPS’s capability. The newly emerging novel approaches may partially address issues that have long plagued researchers, such as low potency, poor cell permeability, on-/off-target toxicity, and delivery efficiency. Comprehensive considerations for the rational design of protein degraders and continuous efforts to seek effective solutions are imperative to advance these strategies into clinical medications.

Introduction

While downregulating protein function through small molecule inhibition has shown great clinical success, limitations such as acquired drug resistance, unavailability for ‘undruggable’ proteins, and relatively high dosages remain challenging. In recent years, disrupting protein function by target protein degradation (TPD) has demonstrated an attractive alternative strategy for drug discovery combating various diseases. The ‘event-driven’ TPD presents the potential to overcome the relevant limitations posed by ‘occupancy-driven’ inhibition, owing to its capability to fundamentally eliminate the pathogenic proteins.

TPD is a concept to achieve the degradation of target proteins via hijacking the inherent recycling system by engineered molecules. The ubiquitin-proteasome system (UPS) and lysosomal system are two primary and complementary protein recycling mechanisms in human cellular homeostasis [1]. UPS degrades relatively small, intracellular, soluble, and short-lived proteins, while lysosomal system prioritizes larger-sized, extracellular, membrane proteins and other components, such as nucleic acids, lipids, and even damaged organelles [2].

Molecular Glue (MG) and PROteolysis Targeting Chimera (PROTAC) are two major strategies to degrade protein of interest (POI, or neosubstrate) via UPS. MG degraders are monovalent molecules that form a new interface with E3 ligase to enhance the interaction of neosubstrate, thereby triggering further degradation (Figure 1A). The earliest MGs were found in plant hormones [3–5], while the largest class of current MGs are derivatives of immunomodulatory imide drugs (IMiDs) represented by thalidomide, lenalidomide, and pomalidomide [6–8]. Notably, most reported MGs are discovered serendipitously; however, novel MGs from rational design approaches are emerging. Some of the top candidates for MGs are currently in clinical trials (Table 1).

Figure 1.

Mechanisms of neosubstrate or POI degradation via UPS by MGs (A) and PROTACs (B).

Table 1.

Representative MGs advanced into clinical trials.^a

Name	Company	Target	E3 ligase	Highest Phase	Indications (NCT Identifier)
Thalidomide	BMS	IKZF1/3	CRBN	Marketed	Multiple Myeloma, Leprosy
Lenalidomide	BMS	IKZF1/3 CK1α	CRBN	Marketed	Multiple Myeloma, Myelodysplastic Syndromes
Pomalidomide	BMS	IKZF1/3	CRBN	Marketed	Multiple Myeloma
Iberdomide (CC-220)	BMS	IKZF1/3	CRBN	III	Relapsed or Refractory Multiple Myeloma (NCT04975997)
Mezigdomide (CC-92480)	BMS	IKZF1/3	CRBN	III	Relapsed or Refractory Multiple Myeloma (NCT05519085)
Avadomide (CC-122)	BMS	IKZF1/3	CRBN	II	Chronic Lymphocytic Leukemia (NCT02406742)
Indisulam (E7070)	Eisai	RBM23/39	DCAF15	II	Acute Myeloid Leukemia, Myelodysplastic Syndromes (NCT01692197)
E7820	Eisai	RBM39	DCAF15	II	Acute Myeloid Leukemia, Myelodysplastic Syndromes (NCT05024994)
CC-99282	BMS	IKZF1/3	CRBN	I/II	Non-Hodgkin Lymphoma, Diffuse Large B Cell Lymphoma (NCT03310619)
CC-90009	BMS	GSPT1	CRBN	I/II	Acute Myeloid Leukemia (NCT04336982)
CFT7455	C4	IKZF1/3	CRBN	I/II	Non-Hodgkin Lymphoma, Multiple Myeloma (NCT04756726)
KPG-818b	Kangpu	IKZF1/3	CRBN	I/II	Systemic Lupus Erythematosus (NCT04643067)
MRT-2359b	Monte Rosa	GSPT1	CRBN	I	Small Cell Lung Cancer, Neuroendocrine Cancer, etc. (NCT05546268)
CC-91633b	BMS	CK1α	CRBN	I	Acute Myeloid Leukemia, Myelodysplastic Syndromes (NCT04951778)
DKY709	Novartis	IKZF2/4	CRBN	I	Nasopharyngeal Carcinoma, Colorectal Cancer, etc. (NCT03891953)
BTX-1188b	BiotheryX	IKZF1/3 GSPT1	CRBN	I	Advanced Solid Tumor, Non-Hodgkin Lymphoma, (NCT05144334)
ORM-5029b	Orum	GSPT1	Undisclosed	I	HER2-Expressing Advanced Breast Cancer (NCT05511844)
KPG-121b	Kangpu	IKZF1/3 CK1α	CRBN	I	Castration Resistant Prostate Cancer (NCT03569280)
TQB3820b	Chia Tai-Tianqing	IKZF1/3	CRBN	I	Hematological Malignancies (NCT05020639)

^aData collected from www.clinicaltrials.gov by Dec 31, 2022.

^bStructures are undisclosed.

PROTAC degraders (hereafter as PROTACs) are heterobifunctional molecules that spatially induce E3 ligase to approach POIs, thereby tagging POIs with ubiquitin for UPS degradation (Figure 1B). Since the first PROTAC was revealed by the Crews’ group in 2001 [9], the development of PROTACs has sprung up rapidly in the past two decades. Some top candidates for PROTACs are currently being investigated in clinical trials (Table 2). Notably, MGs and PROTACs are highly related to each other, as IMiDs are widely used as the E3 ligands in PROTAC design. On the other hand, some PROTACs were recently repositioned as MGs [10].

Table 2.

Representative PROTACs advanced into clinical trials.^a

Name	Company	Target	E3 ligase	Highest Phase	Indications (NCT Identifier)
Vepdegestrant (ARV-471)	Arvinas	ER	CRBN	III	Advanced or Metastatic Breast Cancer (NCT05654623)
Bavdegalutamide (ARV-110)	Arvinas	AR	CRBN	II	Metastatic Castration Resistant Prostate Cancer (NCT03888612)
ARV-766	Arvinas	AR	VHL	I/II	Metastatic Castration-Resistant Prostate Cancer (NCT05067140)
CFT8634	C4	BRD9	CRBN	I/II	Locally Advanced or Metastatic SMARCB1-Perturbed Cancers (NCT05355753)
CFT1946b	C4	BRAFV600X	CRBN	I/II	Solid Tumors, Melanoma, etc. (NCT05668585)
RNK05047b	Ranok	BRD4	Undisclosed	I/II	Advanced Solid Tumor, Diffuse Large B Cell Lymphoma (NCT05487170)
AR-LDD (CC-94676)b	BMS	AR	CRBN	I	Metastatic Castration-Resistant Prostate Cancer (NCT04428788)
DT2216	Dialectic	BCL-2/xL	VHL	I	Solid Tumor, Hematologic Malignancy (NCT04886622)
FHD-609b	Foghorn	BRD9	Undisclosed	I	Advanced Synovial Sarcoma, Advanced SMARCB1-Loss Tumors (NCT04965753)
KT-474b	Kymera	IRAK4	CRBN	I	Atopic Dermatitis, Hidradenitis Suppurativa (NCT04772885)
KT-413b	Kymera	IRAK4	CRBN	I	Non-Hodgkin Lymphoma, Diffuse Large B Cell Lymphoma (NCT05233033)
KT-333b	Kymera	STAT3	Undisclosed	I	Non-Hodgkin Lymphoma, Peripheral T-cell Lymphoma, etc. (NCT05225584)
NX-2127b	Nurix	BTK/IKZF3	CRBN	I	Chronic Lymphocytic Leukemia, Small Lymphocytic Lymphoma, etc. (NCT04830137)
NX-5948b	Nurix	BTK	CRBN	I	Chronic Lymphocytic Leukemia, Small Lymphocytic Lymphoma, etc. (NCT05131022)
BGB-16673b	Beigene	BTK	CRBN	I	B-cell Malignancy, Non-Hodgkin Lymphoma, etc. (NCT05006716)
GT20029b	Suzhou Kintor	AR	CRBN	I	Acne Vulgaris, Androgenetic Alopecia (NCT05428449)
HP518b	Hinova	AR	CRBN	I	Metastatic Castration-Resistant Prostate Cancer (NCT05252364)
HSK29116b	Haisco	BTK	CRBN	I	Relapsed/Refractory B-Cell Malignancies (NCT04861779)
LNK01002b	Lynk	Ras GTPase	CRBN	I	Acute Myeloid Leukemia, Primary Myelofibrosis, etc. (NCT04896112)
ASP3082b	Astellas	KRasG12D	Undisclosed	I	Pancreatic Cancer, Colorectal Cancer, etc. (NCT05382559)

^aData collected from www.clinicaltrials.gov by Dec 31, 2022.

^bStructures are undisclosed.

Lysosome-mediated degradation is achieved via two major pathways, including autophagy-lysosome and endosome-lysosome pathway, among which the former is responsible for cytoplasmic components and the latter induces the degradation of extracellular components. Altogether, lysosomes are responsible for the degradation of targets in the final stage. Lysosome-mediated TPD is somewhat complementary to UPS and has the potential to overcome the limitations of UPS-mediated TPD. Inspired by therapeutic applications of MGs and PROTACs, novel lysosome-based approaches have been developed, such as LYsosome-TArgeting Chimera (LYTAC), AUtophagy-TArgeting Chimera (AUTAC), AUTOphagy-TArgeting Chimera (AUTOTAC), and AuTophagosome-TEthering Compound (ATTEC). Notably, these technologies are currently at an early stage of drug development, with many challenges, such as low in vivo efficacy and on-target toxicity of autophagy receptor binding remaining to be addressed.

In this review, we summarize recently developed novel approaches for TPD based on both UPS and lysosomal system. We hope to highlight key advances that reflect current progress and provide useful perspectives on the challenges and opportunities for TPD-based drug discovery.

1. Molecular Glues (MGs)

MGs act as a ‘glue’ that targets the protein-protein interface resulting in multiple biological responses, such as protein dimerization [11] and protein-protein interaction (PPI) stabilization [12]. The first example of MG was introduced in the 1990s [13]. In the TPD field, typical MGs refer to small molecules that first interact with the E3 ligase, thereby triggering subsequent UPS-based degradation (Figure 1A). Early MGs were discovered serendipitously; however, new discoveries through screening and rational design approaches have also been reported. Inspired by the novel mechanism, emerging MGs have been investigated recently as potential therapeutic agents for treating diseases due to their adhesion ability as a bridge between proteins and promising drug-like properties.

1.1. Plant hormones

Multiple plant hormone signaling pathways are regulated by UPS [14]. Retrospective studies demonstrated that these plant hormones functioned as MGs, exemplified by auxin, jasmonates, and gibberellin. Auxin (1, Figure 2) regulates gene expression through promoting E3 ligase Skp1-cullin 1-F-box (SCF)-catalyzed degradation of the Aux/IAA transcription repressors [15]. A ternary complex co-crystal of auxin, transport inhibitor response 1 (TIR1), and Aux/IAA disclosed the mechanism, by which auxin enhanced the TIR1-substrate interactions as an MG, thereby inducing the degradation of Aux/IAA [3]. Jasmonates are reported to regulate the degradation of jasmonates ZIM-domain (JAZ) protein family in response to host immunity and stress response by binding to SCF substrate coronatine insensitive 1 (COI1) [16]. A co-crystal of JA-Ile (2, a jasmonate derivative) conjugate demonstrated the promotion of physical interaction between COI1 and JAZ1 [4]. Gibberellin was originally identified as a mycotoxin. Model studies of gibberellin signaling in plants suggest that it functions as an MG [5]. Through binding to gibberellin A3 (GA₃, 3), gibberellin insensitive dwarf 1 (GID1) is induced to undergo a conformational switch, which in turn promotes the recognition of transcriptional regulator DELLA proteins (a SCF substrate) by SCF^SLY1/GID2, thereby being ubiquitinated and degraded [17].

Figure 2.

Chemical structures of representative MGs 1–30.

Plant hormones represent the early examples of MGs, although they were serendipitously discovered. Nevertheless, plant hormone-inducible degron system was demonstrated to transplant into nonplant cells that lack corresponding signaling responses, allowing rapid and reversible degradation of neosubstrates [18]. These discoveries demonstrate the feasibility of MGs as general tools for targeting degradation and encourage researchers to seek universal MG modality in humans for therapeutic benefits.

1.2. Immunomodulatory imide drugs (IMiDs)-based MGs

IMiDs-based MGs are the most established MGs. Thalidomide was first approved for anxiety but was later withdrawn due to teratogenicity [19]. Continued research has promoted thalidomide (4) to be launched again for leprosy and multiple myeloma. However, the mechanism of action (MoA) of thalidomide has long remained unclear. Until 2010, it was demonstrated that thalidomide initiates its antimyeloma activity and teratogenicity by binding with E3 ligase cereblon (CRBN) [6]. More studies demonstrated CRBN is the primary target of two other IMiDs, lenalidomide (5) and pomalidomide (6) [7,8].

CRBN is a substrate receptor of E3 complex Cullin-RING ligase 4 (CRL4^CRBN) responsible for neosubstrate recognition and degradation by proteasome. Thalidomide-like IMiDs promote the degradation of Ikaros family zinc finger protein 1/3 (IKZF1/3), which is associated with antitumor and immunomodulatory effects [20]. Moreover, other neosubstrates containing zinc finger (ZF) domains are also preferentially recruited and degraded through IMiDs-dependent pathways, such as casein kinase 1α (CK1α), G1 to S phase transition 1 (GSPT1), zinc finger protein 91 (ZFP91), etc. [21,22]. Inspired by the successful marketing of thalidomide (1998), lenalidomide (2006), and pomalidomide (2013), more IMiDs-based MGs with safety profiles and therapeutic potential have been developed in recent years.

1.2.1. IMiDs-based MGs from structure-based drug design (SBDD)

SBDD is an effective strategy for discovering new IMiDs-based MGs. Following this guide, a representative lenalidomide derivative CC-885 (7) exhibited potent anti-proliferative activity through CRL4^CRBN-induced degradation of GSPT1[23]. Further optimization resulted in CC-90009 (8), a more selective GSPT1-targeting MG, which is currently under phase I clinical trial against acute myeloid leukemia (AML) [24]. 8 was the first MG identified through a rational design. Currently, several next-generation IMiDs-based MGs are under clinical trials (Table 1). Notably, most IMiDs-based MGs target IKZF1/3, indicating the potential of IKZF1/3 degradation for antitumor therapy. However, neosubstrates like CK1α, GSPT1, and IKZF2/4 are also included.

Novel IMiDs-based MGs are excitingly emerging. Fischer’s group reported two novel compounds, ALV1 (9) and ALV2 (10), that contain a more flexible anilinomaleimide core instead of phthalimide [25]. It enabled 9 and 10 to accommodate the key histidine residue in IKZF2, thereby increasing their binding affinity to efficaciously recruit CRBN to IKZF2 for degradation. In particular, 10 exhibited a relative selectivity for IKZF2 over IKZF1/3, indicating the possibility of developing degradative therapeutics targeting IKZF2 through a rational design approach.

1.2.2. IMiDs-based MGs from screening

Kaelin’s group reported a target-driven, E3-independent gain-of-signal (‘up’) assay for identifying protein degraders [26]. The up-assay converts downregulated protein levels into upregulated export signals, thereby avoiding the shortages of classic loss-of-signal (‘down’) assay, such as poor signal-to-noise ratios, narrow dynamic ranges, and false positives [27]. Specifically, they fused deoxycytidine kinase (DCK) and IKZF1, and the degradation of the fusion protein can block the conversion of unnatural nucleoside 2-bromovinyldeoxyuridine (BVdU) into a toxic substance. With the ‘up’ cell viability as the output signal, this group screened and identified a new potent IMiDs-based IZKF1-targeting MG, MI-2–61 (11). The same method was also used to discover spautin-1 (12) as a new scaffold that degrades IZKF1 via a CRBN-mediated pathway. However, 12’s direct target remains unknown.

1.2.3. IMiDs-based MGs from PROteolysis Targeting Chimera (PROTAC) reposition

IMiDs are widely utilized as CRBN ligands for PROTAC design. However, sometimes whether these PROTACs work as true ‘PROTAC’ are less questioned. Recently, some PROTACs have been identified to function through MG mechanism instead. Wang’s group reported a new PROTAC-like molecule, MG-277 (13), which resulted from simple modifications of MD-222 (a bona fide MDM2 PROTAC) [10]. 13 displayed high potency against acute leukemia cell lines, but only slightly induced MDM2 degradation and failed to activate wild-type (wt) p53. The inhibition by MG-277 was MDM2-independent but did require CRBN binding. Further proteomic analysis demonstrated that the target was GSPT1, thus repositioning 13 as an MG. This is the first exciting example that simple modification may convert a PROTAC to MG.

Similarly, Winter’s group identified GSPT1 as an off-target for a series of PROTACs based on promiscuous kinase inhibitors sunitinib and PHA665752. MI-389 (14) displayed pronounced efficacy in leukemia cell lines; however, no kinase degradation was observed, indicating its MG role [28]. Chen’s group reported a PROTAC-like molecule, SJPYT-195 (15), which incorporated SPA70 derivative, a specific inverse agonist of pregnane X receptor (PXR), into thalidomide [29]. 15 reduced PXR level; however, this results from CRBN-mediated GSPT1 degradation. Rao’s group reported a dual-mechanism degrader, GBD-9 (16), which can concurrently induce degradation of BTK and GSPT1 by merging the PROTAC and MG strategies [30]. The anti-proliferative effect of 16 was superior to each single degraders L18I (BTK PROTAC) and CC-90009, but equal to the effect of their combination. For the first time, this study provides insights into the design of degraders that integrate the two mechanisms of MG and PROTAC.

Retrospective studies repositioned thalidomide, lenalidomide and pomalidomide as MGs. SBDD of IMiDs and screening generate various IMiDs-based MGs, representing the largest class of current MGs in development and clinical trials. The structural diversity of IMiDs-based MGs in clinical/preclinical studies indicates a strong structural tolerance of MGs in target binding, while the targets are mostly limited to IZKF1/3 and GSPT1. Whether IMiDs-based can be applied to the degradation of other targets through rational design is noteworthy for further investigations in the future.

1.3. Aryl-sulfonamide MGs

Another type of MG besides IMiDs-based MGs is aryl-sulfonamides, exemplified by indisulam (17). Indisulam has been reported to show potent anticancer activity against several hematological tumors, but the MoA and primary targets of indisulam have remained unclear [31]. In 2017, the research groups of Nijhawan’s and Owa’s first and independently reported the role of indisulam as MG in inducing RNA binding motif protein 39 (RBM39) and coactivator of activating protein-1 and estrogen receptor α (CAPERα) degradation [32,33]. A further retrospective study demonstrated indisulam hijack E3 ligase DDB1 and CUL4 associated factor 15 (DCAF15) to degrade RBM39 and RBM23 [34]. Indisulam was found to function as an MG to facilitate the formation of a new surface of DCAF15 to recruit RBM39, rather than allosterically modulating them, nor by stabilizing weak DCAF15-RBM interactions. The conserved sulfonamide moiety provides the key interactions [35]. Specifically, the central sulfonyl oxygens form interactions with backbone amide from DCAF15 residues, while the nitrogen forms water-mediated H-bonds with RBM39 side chains. Notably, aryl-sulfonamides showed weaker affinity (>50 μM) when binding to DCAF15 compared with IMiDs (100~200 nM), but this deficiency was compensated by the nearly doubling the interfacial surface area between DCAF15 and RBM39. Three analogs with the same central sulfonamide, tasisulam (18), E7820 (19), and chloroquinoxaline sulfonamide (CQS, 20) were also reported as MGs to induce RBM23/39 degradation via the DCAF15-dependent pathway [32,33,36].

In 2020, Winter’s group reported a strategy to find novel MGs through rational discovery via scalable chemical profiling [37]. Through screening on hyponeddylated cellular models with broadly abrogated ligase activity, the author identified several scaffolds capable of reprogramming CRL4^DCAF15, among which the aryl-sulfonamide dCeMM1 (21) was able to degrade RBM23/39.

Currently, the discovery of aryl-sulfonamide MGs is highly dependent on serendipity and screening. However, their structural similarities suggest the feasibility of using aryl-sulfonamide scaffolds as templates to design MGs based on DACF15-mediated degradation.

1.4. MGs that induce cyclin K (CCNK) degradation

Recently, the degradation of cyclin K (CCNK) by MGs has received increasing attention. CCNK interacts with multiple cyclin-dependent kinases (CDKs) (e.g., CDK9/12/13) and involves in the induction of processive elongation [38]. CCNK/CDKs are overexpressed in a variety of cancers, and thus degrading CCNK for therapeutic benefits is emerging as a promising strategy.

In 2020, Ebert’s group rescreened a database of >4500 clinical/preclinical small molecules to discover the correlations between their cytotoxicity and expression levels of E3 ligase components [39]. They identified (R)-CR-8 (22), an established CDK inhibitor, could act as an MG to degrade CCNK via a unique CRL4^DDB1-mediated pathway. Interestingly, CDK12 served as a drug-induced substrate receptor, although it is not a constitutive E3 ligase component. The solvent-exposed pyridyl moiety of 22 played a key role in inducing the formation of a complex between CDK12-CCNK and DDB1, bypassing the requirement of DCAF and hijacking adaptor protein DDB1 for CCNK ubiquitination and degradation. Moreover, while CCNK was the primary target, CDK12 may also be autoubiquitinated after prolonged exposure to 22.

Through the generic screening strategy reported by Winter’s group (cf. 2.3), they identified three MGs, dCeMM2/3/4 (23/24/25), all of which contained substituted acetamide backbones and induced CCNK degradation [37]. When the MG was added, the binding affinity between DDB1 and CDK12-CCNK was greatly increased, resulting in dimerization between the CDK12-CCNK and E3 complex, thereby positioning CCNK for a more susceptible ubiquitination. However, the structural details remained to be elucidated. Similarly, two analogs HQ461 (26) [40] and NCT02 (27) [41], with substituted acetamide backbones were also independently reported.

Collectively, degradation of CCNK/CDK12 leads to decreased serine 2 phosphorylation of RNA polymerase II and affects the expression of genes involved in DNA damage response, highlighting its potential against various cancers such as metastatic colorectal cancer and lung cancer. The CDK12-DDB1 interface is becoming a hotspot for MG development.

1.5. MGs that induce other neosubstrates degradation

MGs with novel mechanisms for targeting various neosubstrates have also been reported. Koegl’s group developed BI-3802 (28) in 2017, an interaction blocker that induces degradation of oncogenic transcription factor (TF) B cell lymphoma 6 (BCL6) [42]. 28 bound to the Broad-complex, Tramtrack and Bric-à-brac (BTB) domain of BCL6 to strongly induce BCL6-repressed genes expression. In-depth research by Ebert’s group proved that 28 promoted dimerization between two BCL6 monomers, with a mechanism different from direct MG-induced E3-neosubstrate coupling [43]. The E3 ligase SIAH1 responded to this dimerization and triggered further degradation. Furthermore, despite the intrinsic but weak binding of SIAH1 to BCL6, the presence of 28 significantly enhanced the interactions, suggesting that 28 was an MG.

Recently, Nomura’s group reported a general method to rationally design MGs for various neosubstrates [44]. Using the CDK4/6 inhibitor ribociclib as a prototype, they identified a covalent derivative EST1027 (29) that induced the depletion of CDK4 in cancer cells via UPS. By screening the chemical handles appended to ribociclib exit vector, a but-2-ene, 1,4-dione (‘fumarate’) handle was selected as the optimal adaptor to achieve the maximal degradation of CDK4 through covalent binding. The E3 ligase ring finger protein 126 (RNF126) was identified to respond to CDK4 degradation. Furthermore, by identification of the minimal covalent motif required for interaction with RNF126, this handle was successfully transplanted to ligands of other neosubstrates, including BRD4, BCR-ABL and c-ABL, PDE5, AR and AR-V7, BTK, LRRK2, and SMARCA2, showing its potential as a general method for converting noncovalent ligands to covalent MGs.

Based on proteome microarray screening, Tu’s group reported bufalin (30), the main active ingredient of some Chinese medicine, could significantly promote TF E2 factor 2 (E2F2) degradation through an atypical E3 ligase ZFP91-mediated pathway [45]. Abnormal E2F2 expression was highly correlated with poor prognosis in various cancers. However, E2F2 was considered ‘undruggable’ because it lacks a typical ligand-binding domain. In this study, 30 acted as an MG bound to ZFP91 and exhibited a specific binding capacity with E2F2 itself. Structurally, 30 was comprised of androsterone, which was responsible for E2F2 binding, and α-pyrone, which covalently interacted with the Cys349 of ZFP91 through electrophilic addition. This study suggests that as a novel atypical E3 ligase, ZFP91 may potentially induce degradation of various neosubstrates other than E2F2. The pyrone group on bufalin may also provide a general structural moiety for discovering novel covalent MGs.

Early MGs are not only limited in structures (e.g., plant hormones and IMiDs), but also limited to degrading certain targets (e.g., IKZF1/3 and GSPT1). New scaffolds from screening provide us with useful strategies to discover novel MGs based on structural modifications. On the other hand, transplanting specific moieties targeting E3 onto off-the-shelf ligands targeting other neosubstrates has the potential to obtain selective MGs, thereby significantly expanding the scope of neosubstrates and enabling their selective degradation.

2. PROteolysis Targeting Chimeras (PROTACs)

PROTACs have attracted significant research attention since it was first reported by Crews and Deshaies’ group in 2001 [9]. PROTACs are heterobifunctional small molecules consisting of a ligand that binds an E3 ligase, a ligand that binds POI, and an appropriate linker that anchors them. While the first-generation PROTACs are peptide-based, current PROTACs are engineered by smaller ligands for better cell permeability and PK properties. Two major E3 ligase ligands, IMiD and hydroxyproline derivatives, which bind to CRBN and von Hippel-Lindau (VHL), respectively, are widely utilized for PROTAC design. By using high-affinity ligands as the POI warheads, PROTACs have successfully degraded various targets, including epigenetic factors, CDK family targets, fusion proteins, membrane proteins, kinases, etc. [46–48]. In addition to widely used CRBN and VHL, other E3 ligases, such as inhibitors of apoptosis proteins (IAPs) can also be used for target protein degradation. Since 2010, IAP-based PROTACs referred to as specific and nongenetic IAP-dependent protein erasers (SNIPERs) have also made significant contributions in TPD [49]. Compared with conventional inhibitors, PROTACs act catalytically to disrupt protein function by directly reducing the expression levels of pathogenic proteins, thereby extenuating side effects from prolonged and high-dose target occupancy.

Moreover, PROTACs have the potential to convert ‘undruggable’ proteins into druggable targets, thus greatly broadening the application of PROTACs in diseases that lack effective therapeutic modalities, such as Alzheimer’s disease (AD) [50]. However, despite the rapid development of PROTACs significant challenges exist, such as low efficacy when utilized alone, on/off-target toxicity, selectivity for diseased cells, and ‘hook effect’. Various novel PROTACs have recently been developed to address such issues, and new ideas for extending the PROTAC toolbox are also emerging, paving the way for viable PROTAC-based drug discovery.

2.1. Dual PROTACs

Inspired by the success of dual/multiple inhibitors, dual PROTACs have recently emerged. Dual PROTACs are PORTACs that can simultaneously target and induce the degradation of two POIs, resulting in a stronger synergistic effect than single agents. There are two subtypes of dual PROTACs, one contains two POI ligands (type A), and the other has only one POI ligand but can bind to two POIs (type B).

Li’s group reported the first example of type A dual PROTACs utilizing gefitinib (an EGFR inhibitor) and olaparib (a PARP inhibitor) as the POI ligands [51]. Both POI ligands have been successfully combined with CRBN or VHL ligands through efficient, convergent synthetic strategies by using a trifunctional natural amino acid as a star-type core linker. DP-C-1 (31, CRBN-based) and DP-V-4 (32, VHL-based) (chemical structures shown in Figure 3) displayed the best simultaneous degradation profiles among each group, which were found superior to each of mono PROTACs. However, 32 showed weaker anti-proliferative activity in H1229 cells compared to the parent inhibitors, likely owing to its poor PK properties. Examples of type B dual PROTACs have also been reported, such as those of Zheng and Zhou’s group [52,53]. Using ABT-263 (a potent dual BCL-2/xL inhibitor) as the POI ligand, they constructed PZ703b (33), exhibiting high potency in degradation of BCL-xL and inhibition of BCL-2 (but not degradation). The stable BCL-2:33:VHL ternary complex was observed, possibly contributing to the enhanced BCL-2 inhibition. Further structural optimization based on the linker length of 33 yielded 753b (34), achieving the desired dual degradation profile on both BCL-xL and BCL-2, and significantly improved the antitumor activity against BCL-xL/2-dependent Kasumi-1 cells.

Figure 3.

Chemical structures of representative PROTACs 31–42 (A) and 43–54 (B).

Developing dual PROTACs appears to be one of feasible approaches to gain stronger synergistic activity than mono PROTACs. However, type A may be more challenging to be developed as a therapeutic agent due to the synthetic difficulties and overall poor PK properties, while developing type B dual PROTACs is likely to be the future trend of dual PROTAC design with more therapeutic potential.

2.2. Macrocyclic PROTACs

Macrocyclic molecules are often described as containing a ring of 12 or more atoms. Confining molecules in their biologically active conformations via macrocyclization for more stable target binding and higher potency represents an attractive strategy for rational drug design. McCoull’s group first reported a macrocyclic PROTAC in 2018 [54]. Through a series of hit-to-lead structural optimization, a macrocyclic BCL6 inhibitor with superior activity, selectively, and cell permeability was identified as the POI ligand. The CRBN-based PORTAC 35 was then constructed to achieve BCL6 degradation in a dose-dependent manner in several hematological cells. However, 35 only showed the same anti-proliferative activity as BCL6 inhibitors in DLBCL cells, likely due to the nuclear residue of BCL6. Olsen’s group reported macrocyclic PROTACs by employing a cyclic tetrapeptide TpxB^Aoda as the POI ligand [55]. The most potent PROTAC 36 enabled the selective degradation of class I HDACs 1–3 in HEK293T cells, demonstrating that macrocyclic peptides can be processed into cell-permeable PROTACs. Ciulli’s group reported a macrocyclic PROTAC, macroPROTAC-1 (37), by combining a cyclizing linker with MZ1 (a selective BRD4 PROTAC) [56]. Guided by the crystal information of MZ1 in complex with BRD4^BD2 and VHL, a cyclization from VHL ligand to POI ligand/linker junction was realized to lock the conformation in the bound state, confirmed by the co-crystal of BRD4:37:VHL ternary complex. Despite a 12-fold loss of binary binding affinity for BRD4, 37 showed comparable cellular activity to that of MZ1.

Macrocyclic PORTACs theoretically improve the efficiency of PROTACs through conformational locking, but some key challenges remain, such as the difficulty of synthesis and the study of detailed binding kinetics.

2.3. Pre-PROTACs

Pre-PROTAC is a concept that acts as a prodrug, releasing active PROTACs under certain conditions. Since complete depletion of a target protein can be lethal, Pre-PROTACs may effectively avoid global toxicity and only degrade POIs in specific tissue when triggered. To achieve this goal, several groups have independently designed novel Pre-PROTACs that can be triggered through optical spatiotemporal control, namely photoswitchable PROTACs and photocaged PROTACs. Moreover, Pre-PROTACs triggered by other conditions, such as cellular oxygen species levels, are also reported.

2.3.1. Photoswitchable PROTACs

In 2019, Carreira’s group reported the first photoswitchable PROTAC [57]. By inducing an ortho-F4-azobenzene linker, they constructed a light-sensitive PROATC, photoPROTAC-1 (38), to realize light-control degradation of BRD2. The azobenzene moiety acted as a switch to transform 38 into an active azo-trans-isomer upon irradiation at 415 nm, after which the linker length was adjusted to form a stable ternary complex, leading to subsequent BRD2 degradation. Notably, this transformation was reversible. When irradiated at 530 nm, an inactive trans-to-cis transformation occurred, thereby eliminating POI degradation. Furthermore, persistent degradation could be achieved without continuous irradiation due to the bistable nature of the azobenzene switch. Jiang’s group reported Azo-PROTAC-4C (39), which was derived from dasatinib (a BCR-ABL inhibitor) by installing azobenzene onto CRBN ligand [58]. Under 361 nm ultraviolet A (UVA), it induced the degradation of ABL and BCR-ABL in K562 cells, while UVC (200~280 nm) irradiation caused partial trans-to-cis conversion. Trauner’s group designed a series of photoswitchable PROTACs by engaging azobenzene or diazocine groups onto different attachment sites of published PROTACs (dBET1 and dFKBP-1) [59]. Under irradiation at 390 nm, PHOTAC-I-3 (40) induced degradation of BRD2/3/4 in RS4;11 cells, while PHOTAC-II-5 (41) induced FKBP12 degradation.

The photoswitching strategy enables reversible conversion of PROTAC active states under different optical conditions, indicating that precise control of PROTAC functions can be achieved through ‘color dosage’. However, current photoswitch incorporation mainly focuses on E3 ligands and linkers, while diverse POI ligands need to be attempted in the future.

2.3.2. Photocaged PROTACs

Another Pre-PROTAC strategy through optical control is photocaged PROTACs, which are irreversible. In 2019, Pan’s group first reported photocaged PROTACs based on dBET1 by installing a 4,5-dimethoxy-2-nitrobenzyl (DMNB) caging group either onto the short amide arm extending from POI ligand JQ1 (a BET inhibitor) or CRBN ligand [60]. The photocaged PROTAC was considered inactive dBET1 and could only restore its activity when the caging group was untied. pc-PROTAC1 (42) was successfully cleaved to release active dBET1 under 365 nm light and degraded BRD4 only after 0.3 min of irradiation and induced expected phenotypic changes in zebrafish. Tate’s group used DMNB to construct a caged VHL ligand on MZ1, affording PROTAC 43 [60]. Intracellular activation of 43 was triggered after 1 min of irradiation, followed by BRD4 degradation in a dose-dependent manner, but remained stable in non-irradiated cells. Wei’s group constructed two photocaged PROTACs, opto-dBET1 (44) and opto-ALK (45), on both of which DMNB carbamates are caged on CRBN ligand to induce BRD3/4 and NPM-ALK degradation in cells, respectively [61]. Deiters’ group installed two new photolabile caging groups, diethylamino coumarin (DEACM) and 6-nitropiperonyloxymethyl (NPOM), onto VHL and CRBN ligands, affording PROTAC 46 and 47, respectively [62]. Under 365 or 405 nm light, 46 induced the ubiquitination of estrogen-related receptor alpha (ERRα), while 47 degraded BRD4 in cells under 365 nm irradiation.

Photocaged PROTACs provide an irreversible manner of PROTAC release, which somewhat mitigates the global toxicity caused by uncaged PROTACs at high doses. However, prolonged exposure to UV may cause additional side effects, such as DNA damage, and the stability of caged PROTACs in darkness remains a concern when handling them in tissues, both of which should be considered in the future to rationally design photocaged PROTACs.

2.3.3. Pre-PROTACs triggered by oxygen species levels in cells

Oxygen species levels play critical roles in the regulation of cell biology. For cancer cells, a hypoxic microenvironment is necessary for their survival, proliferation, and metastasis, which is significantly distinct from that of normal cells [63]. On the other hand, reactive oxygen species (ROS), which affect cell signaling pathways and homeostasis, are dramatically increased in oxidatively stressed cancer cells [64]. The difference in oxygen species levels between normal and cancer cells makes it a potential trigger for the design of Pre-PROTACs to reduce the off-tissue side effects on normal cells.

Taking advantage of the overexpression of nitroreductase (NTR) in hypoxic cells, Zhu’s group designed an NTR-responsive Pre-PROTAC 48, by incorporating a nitroimidazole caging group on VHL ligand [65]. The caging group was recognized by NTR for cleavage, thereby releasing active 48 to efficiently degrade EGFR and subsequently exert antitumor efficacy. Chen’s group reported several caged ROS-responsive PROTACs using an arylboronic acid as the caging group [66]. The authors investigated the attaching sites on the glutarimide or amine of CRBN ligand and found the latter was superior for caging group cleavage when treated with H₂O₂. Finally, Pre-PROTAC 49 efficiently degraded BRD3 and selectively inhibited the proliferation of T47D cells with higher ROS levels. He’s group recently reported a novel approach that applied light and ROS levels as cascading triggers [67]. By conjugating photosensitizers (PS) to POI ligand, they constructed PS-Degrons that can produce cytotoxic singlet oxygen by visible light, thereby causing unspecific oxidative damage of POI and subsequently recruiting UPS for degradation. The most potent PS-Degron, PSDalpha (50), was synthesized by conjugating triphenylamine benzothiadiazole (PS part) and 17β-estradiol (POI ligand) via an acetylene bond, and it completely degraded estrogen receptor α (ERα) to achieve an excellent anti-proliferative activity in MCF-7 cells.

Pre-PROTACs provide feasible ideas to solve the problem of off-tissue toxicity caused by conventional PROTACs; however, the in vivo efficacy has not yet been validated. In addition, the viability of mechanisms to trigger inactive-to-active conversion may be overwhelmed by practical difficulties, such as the penetration of UV light into the skin. Efforts are still needed to address such problems in the future.

2.4. Antibody (Ab)-PROTACs

Since indiscriminative target degradation may lead to severe on-target toxicity, the precise delivery of PORTACs to desired cells/tissues is becoming a hot area of PROTAC discovery. Antibody-drug conjugates (ADCs) have proven to be an effective drug delivery system with great clinical success. Thus, conjugating PROTACs to antibodies (Ab-PROTACs) may provide attractive delivery options for conventional PROTACs, increasing the therapeutic window.

Ab-PROTACs typically consist of an antibody (Ab) as the cell-targeting moiety, a PROTAC as the payload, and a spacer connecting them. Dragovich’s group first reported an Ab-PROTAC, CLL1–22 (51), obtained by conjugating an anti-C-type lectin-like molecule-1 (CLL1) Ab to VHL ligand of GNE-987 (BRD4 PROTAC) via a novel linker [68–70]. GNE-987 was developed by the same group, and while exhibiting picomolar cellular potency, its in vivo exposure was low. A single intravenous dose of 51 afforded sustained in vivo exposures and stronger potency than GNE-987, resulting in antigen-specific tumor regressions. The same group also reported two Ab-PROTACs by conjugating anti-human EGFR2 (HER2) Ab to two ERα PROTACs with XIAP and VHL ligands, respectively [71]. After solving the problems of self-aggregation and instability upon in vivo administration, HER2–12 (52) and HER2–14 (53) were obtained, both of which were well-delivered to MCF7-neo/HER2 cells and gained anti-proliferative activities. This study investigated the effects of three different ADC spacers and demonstrated the importance of selecting an appropriate spacer for efficient intracellular degrader release. Another HER2-targeting Ab-PROTAC 54 using trastuzumab as the Ab part was also reported [72]. 54 could be easily constructed through a copper-free click chemistry reaction between the azido-PEG spacer connecting to BRD4 PROTAC and alkyne-functionalized trastuzumab. Active PROTAC was released after the linker was hydrolyzed following Ab-PROTAC internalization, thereby inducing catalytic BRD4 degradation only in HER2⁺ cells.

Ab-PROTACs offer an attractive solution for delivering PROTACs into specific cells/tissues to minimize side effects and achieve higher potency. However, like ADCs, Ab-PROTACs may suffer from aggregation, immunogenicity, and instability when administered in vivo, which should be considered during further design and development as therapeutic agents.

It is worth noting that some new types of PROTACs, such as peptide-based PROTACs and nucleotide-based PROTACs are also emerging rapidly. Peptide-based PROTACs consist of protein binding domain (PBD), proteasome targeting motif (PTM), and cell penetrating moiety (CPD). Based on this guideline, Zhang’s group reported a cell-permeable peptide-based PROTAC to degrade α-synuclein, which may have potential for the treatment of Parkinson’s disease (PD) [73]. In addition, peptide-based PROTACs with α-helical structure to gain better proteolytic stability, cellular permeability and PK profiles have also been reported [74,75]. Developing nucleotide-based PROTACs represents another innovative approach to address the degradation of TFs and RNA-binding proteins (RBPs) through the binding of oligonucleotide to target proteins. Hall’s group first reported a single-strand RNA-PROTAC to degrade two RBPs, Lin28A (a stem cell factor) and RBFOX1 (a splicing factor) in two cancer cell lines via VHL-mediated pathway [76]. Further efforts have expanded the POI ligand to double-strand nucleotides [77,78] and nucleic acid aptamers [79]. Collectively, while the aforementioned PROTAC approaches cannot fully describe the current status of PROTAC development, these endeavors suggest the way forward for PROTAC to develop technology to make such molecules more widely applicable, more efficient, less toxic, and easier to deliver.

3. Lysosome-mediated TPD

Degradations by UPS are mainly restricted to relatively small, soluble, intracellular targets, limiting their applications in many diseases, such as neurodegenerative disorders (NDD) and infectious diseases [80]. As another recycling mechanism in human cellular homeostasis, lysosomal system provides a complementary approach to degrade a wider range of components, such as extracellular, membrane, aggregated proteins, and even damaged organelles [2]. Typically, the lysosome-mediated degradation can be divided into four stages: 1) induction of autophagy/endocytosis; 2) nucleation, elongation, and formation of the autophagosome/late endosome; 3) fusion with lysosome; and 4) degradation and recycling of the internalized components. Recently, several lysosomal system-mediated TPD techniques, such as LYsosome-TArgeting Chimera (LYTAC), AUtophagy-TArgeting Chimera (AUTAC), AUTOphagy-TArgeting Chimera (AUTOTAC), and AuTophagosome-TEthering Compound (ATTEC) have been developed (Figure 4), which are expected to break through the limitations of MGs and PROTACs, providing new insights for drug discovery against NDD and infectious diseases.

Figure 4.

Mechanisms of (A) LYTACs, (B) AUTACs, and (C) AUTOTACs and ATTECs induced target degradation via lysosomal system.

3.1. LYsosome-TArgeting Chimeras (LYTACs)

LYTACs are engineered molecules consisting of a POI-targeting Ab and a chemically synthesized glycopeptide ligand that binds to lysosome-targeting receptor (LTR), thereby delivering the targets into endosomes for lysosomal-fusion and degradation.

In 2019, Bertozzi’s group first published a prototype LYTACs, in which cation-independent mannose-6-phosphate receptor (CI-M6PR) was selected as the LTR, and its glycan agonists were screened as the LTR binder [81]. By conjugating different antibodies (e.g., Ab-1/2/3) as warheads to glycopeptide ligands, for example, Poly(M6Pn) to construct M6Pn-LYTACs (55a-c), they achieved the degradation of target proteins including membrane proteins, such as EGFR2, CD71 and PD-L1, and extracellular proteins, such as IgG and APOE4. Recently, the same group reported another two LYTACs, Ctx-tri-GalNAc (56) and Ptz-tri-GalNAc (57), by using triantenerrary N-acetylgalactosamine (tri-GalNAc) as the LTR binder to selectively engage a liver-specific LTR, asialoglycoprotein receptor (ASGPR), thereby degrading EGFR and HER2 in a cell-type-specific manner [82]. The conjugation sites (C-terminus, hinge, and CH1 heavy chain) in antibodies were also investigated to find site-specific LYTACs with improved PK in vivo. Tang’s group independently reported a similar tri-GalNAc-LYTACs (58a-d) system to generate a new class of degraders by conjugation with biotin, Abs or fragments Abs [83]. Both exogenous proteins, such as NeutrAvidin, mouse anti-biotin IgG-647, mouse anti-rabbit IgG-647, and endogenous proteins, such as EGFR, are ablated by ASGPR-mediated cell-specific degradation.

Close to the concept of LYTACs, Spiegel’s group proposed Molecular Degraders of Extracellular proteins through ASGPR (MoDE-As) by utilizing small molecules instead of Abs as the POI ligands [84]. MoDE-As facilitate the formation of a ternary complex between POI and ASGPR on hepatocytes, thereby inducing the endocytosis and degradation of targets through the endosomal-lysosomal pathway. The modularity of MoDE-As (59 and 60) was demonstrated to efficaciously deplete extracellular proteins such as α-DNP Ab and proinflammatory cytokine MIF proteins in vitro and in vivo.

LYTAC is an early concept of degrading targets through the lysosomal system, using Abs as POI ligands, but it may have issues such as low cell permeability and poor PK. However, as small molecules, MoDE-As appear more favorable for drug discovery. The merging of LYTAC and MoDE-A concepts may provide clear guidance for designing engineered molecules that degrade targets through the lysosomal system.

3.2. AUtophagy-TArgeting Chimeras (AUTACs)

Microautophagy/Autophagy regulates cellular homeostasis by degrading and recycling cytoplasmic components, such as protein aggregates, damaged organelles, and invading pathogens [2]. In 2019, Arimoto’s group first demonstrated the capability of cargo degradation by exploiting cellular autophagy machinery, which was termed as AUTAC [85]. AUTACs are heterobifunctional chimeras of a small cargo-targeting warhead linked to a guanine derivative tag that recruits the autophagy system. By conjugation Halo tag (as a covalent warhead) with Cys-S-cGMP (as an autophagy tag), AUTACs were capable of selectively degrading enhanced green fluorescent protein (EGFP), which was mediated by autophagy receptors p62/SQSTM1. Several endogenous proteins, such as MetAP2, FKBP12, and BRD4, with different cargo ligands, were also selectively autophagically degraded by AUTACs 61–63. Furthermore, mitochondria-targeted AUTAC 64 also enabled mitophagy of fragmented mitochondria and restoration of overall mitochondrial function, which may have beneficial effects in Down syndrome therapy.

Ouyang’s group architected a series of new AUTACs by conjugation of JQ1 and ligand of microtubule-associated protein 1A/1B light chain 3 (MAP1LC3) GW5074 with a flexible linker, which the autophagy receptor LC3 can induce for lysosome-mediated degradation [86]. The most potent AUTAC 65 was demonstrated to powerfully degrade BRD4 and effectively inhibit breast cancer cell proliferation.

3.3. AUTOphagy-TArgeting Chimeras (AUTOTACs)

AUTOTACs are newly developed autophagy/lysosome-based degradation technology that simultaneously degrades targets and accelerates cellular autophagic flux. In 2021, Kwon’s group first reported a series of bifunctional small molecules, termed as AUTOTACs [87]. AUTOTACs are composed of three parts, including a target-binding ligand (TBL), an autophagy-targeting ligand (ATL) and a linker. ATL binds to the ZZ domain of p62/SQSTM1, thereby inducing self-oligomerization of p62 in complex with cargo proteins and further autophagic degradation through LC3 interaction. Unlike PROTACs, LYTACs, and AUTACs, ubiquitination is not involved in AUTOTAC-mediated degradation. By applying PHTPP, Vinclozolin M2, and Fumagillin as the TBLs respectively, three AUTOTACs, PHTPP-1304 (66), VinclozolinM2–2204 (67), Fumagilin-105 (68) were constructed by Kwon’s group to induce LC3-mediated autophagy-lysosomal clearance of monomeric proteins including ERβ, AR, and MetAP2 [87]. Notably, not only monomeric proteins, but also aggregation prone proteins, such as misfolded tau, can be efficiently removed by AUTOTAC-mediated degradation approach.

3.4. AuTophagosome-TEthering Compounds (ATTECs)

ATTECs are small molecules that can hijack selective autophagy machinery for degradation of POI through an autophagy receptor-dependent manner, which Lu’s group first proposed in 2020 [88,89]. By screening against a library of >3000 compounds and structural optimization, they discovered four ATTECs, including GW5074 (69), ispinesib (70), and two analogs, AN1 (71) and AN2 (72). They enabled the reduction of mutant huntingtin protein (mHTT) but not wtHTT in cellular assays and Huntington disease (HD)-related animal models via LC3-mediated autophagy. The expanded polyglutamine (polyQ) stretch of mHTT played a critical role in LC3 recognition, expanding the utility of ATTECs to polyQ-containing proteins, such as mutant ATNX3. Similar to AUTOTACs, ATTEC-mediated degradation bypasses the requirement of ubiquitin. Encouraged by the efficacious rescue of disease-relevant phenotypes and excellent brain permeability, ATTECs are expected as a general therapeutic strategy for HD and other diseases caused by polyQ proteins. Moreover, the same group also achieved the degradation of non-protein pathogenic biomolecules (e.g., lipid droplets (LDs)) by the same strategy through LC3-mediated autophagic degradation, which was termed as LD-ATTECs [90].

Sheng’s group reported a series of ATTECs that connected the nicotinamide phosphoribosyltransferase (NAMPT) inhibitor MS2 and LC3-binding ispinesib with changeable linkers [91]. In particular, 73 significantly decreased NAMPT level. Although this ATTEC was found not superior to NAMPT PROTACs, it provided an alternative strategy for NAMPT degradation.

4. Conclusions

In this review, we summarized recent TPD technologies exploiting UPS and lysosomal system. UPS-based approaches include MGs and PROTACs. MGs were first serendipitously discovered in plant hormones, while the largest class of MGs currently in development and clinical trials are IMiDs-based. In addition, novel scaffolds such as aryl-sulfonamides, and new targets such as CCNK, and BCL6 are emerging for MGs. PROTACs have been extensively investigated against numerous targets in the past decade. Still, several critical issues remain, such as low efficacy of mono PROTAC, on/off-target toxicity, selectivity for diseased cells, poor delivery efficiency, and the lack of guiding principles to select suitable E3 ligases, which significantly hinder their clinical translation. Recently developed approaches such as dual PROTACs, macrocyclic PROTACs, Pre-PROTACs, and Ab-PROTACs provide possible solutions to address these challenges. Lysosome-mediated approaches exemplified by LYTACs, AUTACs, AUTOTACs, and ATTECs have been recently investigated as a complement to UPS-mediated TPD. These technologies enable protein degradation for a wider range of targets unavailable for UPS, such as aggregated proteins, extracellular proteins, and even damaged organelles. However, Lysosome-mediated TPD is still at the early stage of drug development, suggesting that more research efforts are imperative to verify their generalizability and translate them to in vivo applications.

5. Expert Opinion

TPD has attracted increasing attention due to its advantages over inhibition in degrading ‘undruggable’ proteins and avoiding high dosages. Strategies have been developed by utilizing two primary processes in human cellular homeostasis, MGs and PROTACs, and the recently proposed lysosome-mediated TPD. Typical MGs (e.g., IMiDs) first form a new interface with E3 ligase, enhancing the binding of neosubstrates for degradation, but atypical MGs (e.g., CR-8) that first interact with neosubstrates to recruit E3 ligase is also reported [39]. Three approaches have been explored for generating MGs: serendipitous discovery, screening, and rational design. Serendipitous discovery and screening approaches lack a systematic structure-activity relationship (SAR) for MG design. However, despite current MGs being mostly IMiD-based and target-restricted, structural diversity is increasing, and new scaffolds screened are constantly emerging, promoting the possibility of rationally designing MGs to degrade more neosubstrates. Furthermore, examples of PROTAC repositioning to MG highlight structural tolerance when designing MGs for specific targets, while downsizing to basic structures is desirable for achieving better drug-like properties and avoiding off-target toxicity [10,28,29]. Transplanting functional moieties to off-the-shelf ligands is another effective strategy for discovering MGs to selectively degrade multiple targets [44,45]. On the other hand, in-depth research on degradation mechanisms, exemplified by dimerization-induced BCL6 degradation and CDK12-promoted CCNK degradation, is also helpful for developing new MGs [39,43].

Although PROTACs can theoretically be used for degradation of both druggable and ‘undruggable’ targets, current clinical trials are limited to certain drug targets such as AR and BTK, indicating that most of the current PROTACs have limited druggability. Novel PROTAC technologies have emerged continuously over the past decade, with the exciting representative examples discussed in this review. The development of PROTACs can be roughly divided into two directions. One is to solve the existing issues of conventional PROTACs, such as low potency, on/off-target toxicity, poor PK, and delivery efficiency, and some novel approaches summarized in this article partially address such issues. Among them, Pre-PROTAC is one of the most advantageous strategies since it can precisely control the release of PROTAC without affecting its activity. However, current pre-control is mainly triggered by UV light irradiation, which has limited penetration and tissue targeting, and the antitumor activity in vivo has not been verified. Intriguingly, an X-ray control Pre-PROTAC was recently reported as effective in the MCF-7 xenograft model [92]. Other Pre-PROTACs triggered by cellular oxygen species levels suggest the possibility of applying other conditions, such as temperature, magnetic field, ultrasound, etc., as triggers, but still require ingenious conception and design by researchers. Some other strategies, such as covalent PROTACs [93] and trivalent PROTACs [94] are also investigated. The other direction for PROTAC development is to expand the application space of PROTACs, either for expanding the scope of E3 ligases or expanding clinical therapeutic applications beyond cancer. While numerous POIs have been utilized for targeting, the exploration of E3 ligase is limited to a relatively small range, such as CRBN and VHL. Recent progress indicates that more E3 ligases are introduced into PROTACs; however, their efficiency and generality remain to be explored [95]. PROTACs have gradually been applied in fields other than cancer, and breakthroughs have been made in the fields of NDD, such as tau-PROTACs [50], and antibacterial, such as BacPROTACs [96]. Moreover, new PROTAC technologies for real-time detection and tracking the protein degradation in living cells, namely theranostic PROTACs [97], for retaining targeted protein stabilization via deubiquitination pathway, namely DeUBiquitinase-TArgeting Chimeras (DUBTACs) [98], and for inducing essential protein degradation selectively in cancer cells through proximity to overexpressed oncoproteins, namely Regulated Induced Proximity TArgeting Chimeras (RIPTACs) [99] have also begun to emerge and remain to be extensively investigated as potential paradigm-shifting strategies.

Lysosome-mediated TPD strategies are actively pursued, providing an alternative approach to targeted degradation other than via UPS, especially for diseases like NDD and infectious diseases. LYTACs are Ab conjugates, and AUTACs and AUTOTACs are bifunctional molecules like PROTACs, indicating that the PK and bioavailability issues are important factors affecting their clinical translation, and should be optimized prior to the potential clinical application. Small molecule LYTACs (a.k.a. MoDE-As) appear more favorable for drug discovery. Original ATTECs are small molecules, and their PK can be optimized through SAR studies. However, the discovery of ATTECs is currently highly dependent on screening. Interestingly, ATTEC is a reliable strategy to provide LTR ligands (e.g., GW5074) for AUTAC design, which is similar to the function of IMiDs in PROTAC design. LYTACs, AUTACs, AUTOTACs and ATTECs are TPD strategies involving vesicles, and novel chaperone-mediated autophagy targeting platforms that do not require vesicles have also been established [100]. It is worth noting that lysosomal system is more complex and branched degradational than UPS, due to its multiple canonical (e.g., catabolism) and non-canonical (e.g., secretion and trafficking) functions, thereby significantly hindering their clinical transition. Nevertheless, the advantages of lysosome-based TPD technologies inspire researchers to continuously broaden their potential applications in the future.

As a result of TPD technologies, drug discovery for cancer and other diseases is progressing at a remarkable pace. As research deepens, we anticipate that more novel approaches will likely be developed soon, leading to the discovery of viable drug candidates that can be established as new medications for patients.

Figure 5.

Chemical structures of representative LYTACs, AUTACs and ATTECs 55–73.

Article highlights.

• Ubiquitin-proteasome system (UPS) and lysosomal system are two primary human protein homeostasis mechanisms widely exploited to achieve targeted protein degradation (TPD).

• Molecular Glue (MG) and PROteolysis Targeting Chimera (PROTAC) are two major strategies for UPS-based protein degradation, while lysosome-mediated TPD strategies include LYsosome-TArgeting Chimera (LYTAC), AUtophagy-TArgeting Chimera (AUTAC), AUTOphagy-TArgeting Chimera (AUTOTAC), and AuTophagosome-TEthering Compound (ATTEC).

• MGs were first discovered in plant hormones, and the largest class of current MGs in development and clinical trials are immunomodulatory imide drugs (IMiDs)-based.

• Most MGs are discovered serendipitously or from screening, but rational design approaches for novel MGs are emerging.

• The developmental trend of PROTACs is to achieve higher potency, lower on-/off-target toxicity, and more efficient delivery, with novel approaches highlighted by developing dual PROTACs, macrocyclic PROTACs, Pre-PROTACs, and Ab-PROTACs.

• Lysosome-mediated TPD technologies enable the degradation of extracellular proteins, aggregated proteins, damaged organelles, etc., significantly expanding the range of drug targets.

• Newly emerging PROTAC technologies such as theranostic PROTACs, DeUBiquitinase-TArgeting Chimeras (DUBTACs), and Regulated Induced Proximity TArgeting Chimeras (RIPTACs) may offer potential paradigm-shifting strategies.