Proteolysis targeting chimeras (PROTACs) are emerging therapeutics for hematologic malignancies

Anaplastic lymphoma kinase (ALK) is a receptor tyrosine kinase which is activated in many cancers including several hematologic malignancies (e.g., anaplastic large-cell lymphoma (ALCL) and diffused large B cell lymphoma (DLBCL)) and solid tumors (e.g., non-small cell lung cancer (NSCLC)) due to chromosomal translocations, substitution mutations, and gene amplification [49]. Several ALK inhibitors (crizotinib, ceritinib, alectinib, and brigatinib) have been approved for the treatment of ALK-positive NSCLC [50], and some of them are undergoing clinical trials against ALCL and other lymphomas [Identifier: NCT02465060; NCT00939770; NCT03719898]. The efficacy of ALK inhibitors is hindered by the emergence of different resistance mechanisms [50]. Researchers have adopted PROTAC technology to overcome the resistance to ALK inhibitors. The first series of ALK PROTACs were reported by Gray’s group. These PROTACs were very efficient in degrading ALK (DC₅₀ ~ 10 nM in H3122 NSCLC cells) and inhibiting the proliferation of ALK-dependent ALCL and NSCLC cells. However, these PROTACs were not specific to ALK and could not degrade a mutated ALK fusion protein EML4-ALK [51]. At the same time, another group reported two ALK PROTACs (MS4077 and MS4078) that efficiently degraded ALK fusion proteins NPM-ALK and EML4-ALK in SU-DHL-1 ALCL and NCI-H2228 NSCLC cells, respectively, and potently inhibited the proliferation of SU-DHL-1 cells [5]. Another VHL-based ALK PROTAC TD-004 efficiently induced ALK degradation and inhibited the proliferation of SU-DHL-1 and H3122 cells in vitro, and reduced H3122 xenografted tumor growth in vivo [41]. Recently, a VHL-recruiting ALK PROTAC based on brigatinib, named SIAIS117, was found to be more potent than brigatinib in inhibiting the growth of G1202R mutant ALK-expressing 293T cells by inducing G1202R mutant ALK degradation [52]. The PROTACs against ALK have also been briefly discussed in a review by Kong et al. [53].

Resistance to apoptosis plays a crucial role in tumorigenesis and is responsible for resistance to cancer therapies [54]. Therefore, targeting the apoptotic pathway becomes an attractive therapeutic strategy for cancer treatment. B cell lymphoma 2 (Bcl-2) proteins control the intrinsic mitochondria-mediated apoptotic pathway [55, 56]. SMIs targeting the anti-apoptotic Bcl-2 family proteins, including Bcl-2, Bcl-xL, and Mcl-1, have been developed for cancer treatment. Venetoclax (ABT-199), a highly selective inhibitor of Bcl-2, is the first FDA-approved Bcl-2 antagonist for the treatment of various hematologic malignancies including chronic lymphocytic leukemia (CLL) and small lymphocytic lymphoma (SLL) as a single agent, and for acute myeloid leukemia (AML) in combination with chemotherapy [57]. Wang et al. reported a Bcl-2 PROTAC with a DC₅₀ of 3.0 μM in NCI-H23 lung adenocarcinoma cell line [58]; however, neither degradation nor cellular cytotoxicity data was available in hematologic tumor cell lines. Navitoclax (ABT-263), a dual inhibitor of Bcl-2 and Bcl-xL, entered clinical trials in 2006. Unfortunately, the clinically effective dosage of ABT-263 is significantly limited by thrombocytopenia because platelets solely depend on Bcl-xL for survival [59]. Recently, we reported the first example of utilizing PROTAC technology to reduce on-target toxicity of SMIs [4]. DT2216, a representative PROTAC that hijacks the VHL E3 ligase to achieve tissue/cell-selective degradation of Bcl-xL, was found to be more potent against Bcl-xL-dependent tumor cells and less toxic against platelets in comparison to conventional Bcl-xL inhibitors. Moreover, DT2216 potently degraded Bcl-xL protein and inhibited MOLT-4 T-ALL in vitro and in vivo. The discovery of DT2216 provides a proof of principle to utilize PROTAC technology to rescue undruggable targets due to previously unmanageable on-target toxicities. Using a similar strategy, we reported several PROTACs that recruited CRBN or IAPs for Bcl-xL degradation [60‐63]. These Bcl-xL degraders showed significantly reduced cytotoxicity against platelets but improved killing of various tumor cells and senescent cells due to differential expression of these E3 ligases in these cells, indicating that an improved therapeutic window can be achieved by converting a non-selective SMI to a tissue/cell-specific degrader.

Mcl-1 is upregulated in various hematologic malignancies (e.g., multiple myeloma (MM) and AML) and some solid tumors (e.g., hepatocellular carcinoma and NSCLC) [64]. Tremendous efforts in drug discovery programs have generated highly selective and potent Mcl-1 inhibitors with a binding affinity at picomolar levels. Some of them (AMG-176, AMG-397, AZD5991, S64315/MIK665, and ABBV-467) have entered phase I clinical trials [65]. Notably, in late 2019, the FDA and Amgen halted the clinical trials of AMG-176 and AMG-397 due to a “safety signal for cardiac toxicity,” which is likely to be an on-target toxicity of Mcl-1 inhibition. There are reported Mcl-1 PROTACs [45, 58], dMCL1-2 as a representative, which degrades Mcl-1 in OPM2 and MM.1S MM cell lines. It is worth looking into the possibility to design tissue-specific PROTACs to reduce the on-target toxicity of Mcl-1 inhibition in the heart.

B cell lymphoma 6 (Bcl-6) is an emerging oncoprotein and therapeutic target for lymphoma. It is broadly expressed in many lymphomas and plays a significant role in lymphomagenesis by being involved in the formation of germinal centers (GCs) during the humoral immune response [66‐69]. As a transcriptional repressor, Bcl-6 exerts its function by recruiting its corepressors to its N-terminal BTB domain to repress the expression of a broad set of genes involved in the proliferation and survival of GC B cells [70]. By targeting the interactions between the BTB domain of Bcl-6 and its corepressors, a variety of Bcl-6 inhibitors have been developed over the past decade, including peptide mimetics, small molecules, and natural compounds [71]. However, due to the challenge of disrupting PPI, these Bcl-6 inhibitors have suffered from low potency [72] and thus have limited clinical application. Chemically induced degradation of Bcl-6 by small molecules, such as BI-3802, have been found to have stronger effects than Bcl-6 inhibition and thus offer new routes to the development of lymphoma treatments [73]. By degrading Bcl-6 via the UPS, BI-3802 displayed stronger de-repression of target genes than Bcl-6 inhibitors and demonstrated antiproliferative effects in several DLBCL cell lines [73]. However, inducing Bcl-6 degradation through PROTAC strategy has met with limited success. McCoull et al. developed highly potent Bcl-6 inhibitors and converted them into PROTAC molecules by conjugating them with thalidomide [42]. Although the PROTAC molecules could degrade Bcl-6 in a number of cells, they did not improve antiproliferative activity over Bcl-6 inhibitors in DLBCL cells. Further investigation revealed that the weak antiproliferative response is due to the incomplete Bcl-6 degradation [42]. It remains unknown if complete degradation of Bcl-6 can be achieved through the exploration of the chemical space of their PROTAC molecules or through the recruitment of different E3 ligases.

BCR-ABL is an oncogenic fusion protein implicated in the development of chronic myelogenous leukemia (CML) [74]. Tyrosine kinase inhibitors (TKIs, e.g., imatinib, dasatinib, and bosutinib) that inhibit the kinase function of BCR-ABL by competitively binding at the ATP-binding site of ABL are the standard-of-care against BCR-ABL-driven CML [75]. However, TKIs need to be administered continuously for life to treat CML patients likely due to the scaffolding functions of BCR-ABL that activate compensatory signaling pathways to promote the survival of leukemic stem cells (LSCs) [76, 77]. Since PROTACs eliminate both enzymatic and scaffolding functions of targeted proteins by inducing its degradation, they can potentially overcome the abovementioned limitations of BCR-ABL inhibitors. This led to the development of the first BCR-ABL PROTAC by Crews’ group. In this study, a lead PROTAC based on dasatinib recruiting CRBN (DAS-6-2-2-6-CRBN) degraded both ABL and BCR-ABL, whereas one recruiting VHL (DAS-6-2-2-6-VHL) could only degrade ABL. DAS-6-2-2-6-CRBN effectively reduced the viability of BCR-ABL-dependent K562 CML cells with an EC₅₀ of 4.4 nM [78]. In the same year, Naito’s and Kurihara’s groups developed the specific and non-genetic IAP-dependent protein erasers (SNIPERs) against BCR-ABL. The lead SNIPERs, SNIPER(ABL)-2 and SNIPER(ABL)-3, degraded both BCR-ABL and ABL leading to growth inhibition in K562 cells at concentrations of 30 μM or higher [79]. Subsequently, Naito’s group developed a more potent BCR-ABL SNIPER, SNIPER (ABL)-39 that degraded BCR-ABL and inhibited the proliferation of several BCR-ABL-dependent CML cells at lower nanomolar concentrations [80]. Recently, Crews’ group reported the development of BCR-ABL PROTACs based on an allosteric inhibitor (GNF-5) as the warhead linked to the VHL ligand. The lead PROTAC in this study (GMB-475) induced the degradation of both BCR-ABL1 and ABL1 in a concentration- and time-dependent manner and inhibited the proliferation of K562 cells with an EC₅₀ of ~ 1 μM. Interestingly, GMB-475 inhibited the proliferation of imatinib-resistant murine Ba/F3 cells expressing mutant BCR-ABL1 (T315I) with similar potency as the cells expressing wild-type BCR-ABL1 [81]. Recently, another study reported the most potent BCR-ABL PROTAC to date, named SIAIS178. SIAIS178 potently degraded BCR-ABL protein (DC₅₀ = 8.5 nM) and inhibited the proliferation of K562 cells (EC₅₀ = 24 nM), and further efficiently inhibited tumor growth and induced BCR-ABL degradation in K562 xenografts. More importantly, SIAIS178 was able to degrade several mutant forms of BCR-ABL [6]. More recently, light-controllable Azo-PROTACs against BCR-ABL have been reported with high potency in K562 cells [82].

The bromodomain and extra terminal domain (BET) family includes four bromodomain-containing proteins (BRDs): BRD2, BRD3, BRD4, and BRDT. Each BRD contains two bromodomains that serve as “epigenetic readers” to recognize acetylated lysine residues. In vitro and in vivo studies demonstrate that BET inhibitors (BETi), such as JQ1 and I-BET151, can suppress the proliferation of multiple types of tumors [83]. However, acquired resistance limits the efficacy of BETi treatment in clinical trials [84]. In addition, long-term dosing or high concentrations associated with BETi treatment also leads to obvious BRD4 protein accumulation and inefficient c-MYC suppression [85].

Recently, numerous BET PROTACs have been generated and tested in vitro and in vivo. Different E3 ligases including VHL [86‐89], CRBN [43, 85, 90‐93], IAP [94], MDM2 [95], AHR [96], DCAF16 [97], RNF114 [98], and RNF4 [99], were recruited to degrade BET proteins. Most of the reported PROTACs not only degrade BRD4, but also BRD2 and BRD3. However, simultaneous degradation of the three BRD proteins may increase drug toxicities [100‐102]. To improve the specificity of the PROTACs, Ciulli’s group reported the first BRD4-specific PROTAC (ZXH-3-26). However, no cellular antiproliferative efficacy against cancer cells was reported for ZXH-3-26 [103].

In 2015, Bradner’s group reported the first small molecule-based BRD4 degrader, dBET1 [91]. Later, they optimized the linker and found a more potent PROTAC (dBET6) that showed better efficacy against T-ALL compared with JQ1 [43]. In the same year, Crews’ group reported a CRBN-based BET PROTAC, ARV-825, by utilizing a more potent BETi (OTX015) as the warhead for BRD4 [85]. In 2016, a VHL-based PROTAC (ARV-771) that also used OTX015 as the BRD4 warhead was generated [86]. Later, ARV-825 and ARV-771 were tested widely and displayed potent in vitro and in vivo efficacies against AML [104, 105], post-myeloproliferative neoplasm secondary AML [106, 107], MM [108], mantle cell lymphoma (MCL) [8], and DLBCL [109].

In 2018, Wang’s group reported two more potent CRBN-based BET PROTACs, BETd-260 [92] and QCA570 [93]. BETd-260 exerted high potency to rapidly degrade BRD2/3/4 and inhibit the growth of RS4-11 B cell acute lymphoblastic leukemia (B-ALL) cells in vitro and in a mouse xenograft model [92]. QCA570 is the most potent BET PROTAC that displays picomolar potency against human acute leukemia cell lines in vitro and in vivo [93]. In 2019, Crews’ group reported an MDM2-based BRD PROTAC (A1874), which not only degrades BRD4, but also stabilizes p53, and exhibited high anti-proliferative effect in several tumor cell lines including BL and myeloid leukemia [95]. Very recently, TD-428, a new BRD4 PROTAC comprised of a novel CRBN binding moiety (TD-106) and the BET binding moiety JQ1, was reported [110]. TD-428 possesses two unique properties: high specificity and lacking the Hook effect. Compared with ARV-825, TD-428 only induces moderate IKZF1/3 degradation at higher concentrations. Jiang et al. reported a new CRBN-based PROTAC (compound 15) which selectively degrades BRD2/4 over BRD3. Compound 15 suppresses the proliferation of different AML cells (MV4-11 and MOLM-13) as well as solid tumor cell lines in the NCI-60 cell line panel [111]. Interestingly, in a series of previously reported VHL-based PROTACs derived from a more potent BETi, tetrahydroquinoline (I-BET726), only selectively induce BRD3/4 degradation without any effect on BRD2, but still possess the anti-proliferation effect against AML cells [89].

Bruton’s tyrosine kinase (BTK) is involved in constitutively active B cell receptor (BCR) signaling in various B cell malignancies CLL, DLBCL, and MCL, and thus, is an important therapeutic target in these malignancies [112]. Ibrutinib, an irreversible covalent inhibitor of BTK, is an approved drug to treat CLL and other BTK-driven cancers. Although most of the CLL patients on ibrutinib therapy show durable responses, the resistance to ibrutinib emerges due to a cysteine to serine mutation (C418S) at the ibrutinib binding site in BTK [113, 114]. Therefore, researchers have aimed to employ PROTAC technology to overcome the resistance. Using a quantitative proteomic approach, Gray’s group reported the degradation of multiple kinases including BTK by a multi-kinase degrader, TL12-186. They also reported two selective BTK degraders, one based on a promiscuous inhibitor (bosutinib) and another based on a selective BTK inhibitor (RN486). The RN486-based degrader, named DD-04-015, efficiently and selectively degraded BTK and produced a similar antiproliferative activity as RN486 in BTK-dependent TMD8 DLBCL cells. However, the ability of these PROTACs to degrade mutant forms of BTK was not evaluated [115]. Later, MT-802, a lead BTK PROTAC developed by Crews’ group, was shown to be equally effective in degrading both wild-type and C418S mutated BTK. MT-802, but not ibrutinib, was found to reduce activated-BTK levels in patient-derived CLL cells harboring the C481S mutation [116]. Using a similar strategy, Rao’s group reported a BTK degrader, named P13I, which efficiently degraded both wild-type and C418S BTK with DC₅₀ values of 10 nM and 28 nM, respectively. More importantly, P13I more effectively reduced the viability of BTK-dependent wild-type and C418S BTK-expressing HBL-1 DLBCL cells compared to ibrutinib [117]. Subsequently, several potent BTK PROTACs aimed at targeting mutant BTK in ibrutinib-resistant B cell malignancies with improved cellular activities and possessing in vivo potencies were reported by different groups [7, 44, 118‐120]. Recently, a light-controllable PROTAC against BTK, named pc-PROTAC3, was reported to efficiently degrade BTK in Ramos Burkitt’s lymphoma (BL) cells upon light exposure. pc-PROTAC3 was designed by attaching a bulky photo-removable group, 4,5-dimethoxy-2-nitrobenzyl (DMNB), to the imide nitrogen of a previously described BTK PROTAC MT-802 [23].

CDK6 is a member of cyclin-dependent kinases (CDKs) and regulates G1-S phase transition and drives cell proliferation by forming complexes with D-type cyclins to phosphorylate the tumor suppressor retinoblastoma (Rb) [121, 122]. Multiple types of leukemia cells are dependent on CDK6 for survival, including mixed-lineage leukemia (MLL) [123] and Philadelphia chromosome-positive ALL (Ph⁺ ALL) [9]. However, selectively targeting CDK6 with ATP-competitive SMIs has been challenging due to the presence of an identical ATP-binding domain in its close-related homology CDK4 [9]. Moreover, ATP-competitive inhibitors cannot disrupt the scaffolding functions of CDK6, which is important for the transcription function of CDK6 [124, 125]. Soon after the degradability of CDK4/6 and other CDK family kinase was demonstrated by Huang et al. with a multi-kinase PROTAC degrader TL12-186 [115], several groups have developed selective CDK6 PROTAC degraders by converting CDK4/6 dual inhibitor (palbociclib) to PROTAC molecules [9, 121, 126]. For example, BSJ-03-123, a CRBN-based CDK6 degrader, was reported to selectively degrade CDK6 but not CDK4 in AML cell lines because it can form a ternary complex with CDK6 and CRBN but not with CDK4 and CRBN [121]. Dominici et al. demonstrated that CDK6 degradation by PROTACs is more effective than treatment with the dual CDK4/6 inhibitor palbociclib in suppressing Ph⁺ ALL in mice [9], because PROTACs disrupted both the kinase-dependent and independent activities of CDK6. YX-2-107, a CRBN-recruiting PROTAC, degraded CDK6 selectively in Ph-like and MLL-rearranged cell lines, and inhibited their proliferation. Moreover, YX-2-107 is effective in a Ph⁺ ALL xenograft mouse model. However, the limited pharmacokinetic (PK) properties (e.g., short half-life) of YX-2-107 warrant its further optimization.

FMS-like tyrosine kinase-3 (FLT-3) belongs to the class III receptor tyrosine kinases, which is activated by a FLT-3 ligand and plays a major role in the regulation of hematopoiesis [127]. Mutations in the FLT-3 gene have been found in patients with ALL (1–3%), myelodysplasia (5–10%), and AML (15–35%), making it one of the most frequently mutated genes in hematologic malignancies [128]. Mutations in the FLT-3 gene, such as the most common internal tandem duplication (ITD), result in its constitutive activation in these malignancies. Thus, FLT-3 has been emerged as a therapeutic target for hematologic malignancies. Several FLT-3 inhibitors have been tested in clinical trials and two (midostaurin and gilteritinib) of them have been approved by FDA for the patients with FLT3-mutated AML [129]. However, clinical response duration to FLT3 inhibitor treatment remains short partially due to the development of secondary resistance [130]. Achieving a complete and sustained inhibition of FLT-3 ITD signaling to maintain effective clinical response requires a high dose of inhibitors, which leads to off-target toxicities [131].

In an early study, using a CRBN-recruiting PROTAC based on a promiscuous kinase inhibitor to target multiple kinases failed to significantly degrade FLT-3, likely due to its poor cell permeability [115]. In contrast, a VHL-recruiting FLT-3 PROTAC from Crews’ lab potently induced FLT-3 ITD degradation in MV4-11 and MOLM-14 AML cell lines in vitro [10]. They confirmed the ability of this FLT-3 PROTAC to degrade FLT-3 in MV4-11 mouse xenografts. This study demonstrated that conversion of an FLT-3 inhibitor into a PROTAC can result in the degradation of FLT-3 ITD. However, the in vitro anti-proliferative activity of the FLT-3 PROTAC varies among AML cells compared to the FLT-3 inhibitor that was used to develop the PROTAC. For example, the PROTAC exerted a significantly better activity against MV4-11 and wild-type MOLM-14 AML cells, but was less potent in OCI-AML3 cells, than the FLT-3 inhibitor, which is likely attributed to the extent of kinase engagement/selectivity of the PROTAC [10]. In addition, both the FLT-3 PROTAC and inhibitor were ineffective against other types of hematologic malignancies, such as K562 CML cells, suggesting that the antitumor effect of the FLT-3 PROTAC on different FLT-3 ITD mutated cells is cell-dependent.

Histone deacetylases (HDACs) and histone acetyltransferase (HAT) are essential regulators of lysine acetylation which determines chromatin accessibility during transcription. Dysregulation of HDACs is commonly observed in both hematologic and solid malignancies, and inhibition of HDACs leads to apoptosis, differentiation, and growth arrest in tumor cells. There are 11 isoforms of zinc-dependent HDACs in humans, which can be divided into four classes: class I (HDACs 1, 2, 3, and 8); class IIa (HDACs 4, 5, 7, and 9); class IIb (HDACs 6 and 10); and class IV (HDAC 11) [132].

To date, four HDAC inhibitors (HDACi) have been approved by the FDA for the treatment of lymphomas, leukemias, and MM. Most of the approved drugs have limited isoform selectivity and none of the drug candidates have achieved clinical success for treating solid tumors as a single agent, which is probably due to their inability to achieve high concentrations in solid tumor tissues. Importantly, dose-limiting adverse effects such as cardiac toxicity associated with hERG K⁺ channel activation are hindering their progress in the clinic [133, 134].

By converting an HDACi into an isoform-selective HDAC PROTAC, it is possible to achieve improved potency and reduced off-target toxicity as reported recently [11, 135‐138]. CRBN-based PROTACs recruiting a pan-HDAC inhibitor AB3 or a HDAC6 selective binder nexturastat A resulted in specific degradation of HDAC6. Wu et al. proposed that HDAC6-IKZFs dual degraders would achieve an enhanced anti-myeloma activity based on the synergistic effect of HDAC6 inhibitors and immunomodulatory imide drugs (IMiDs) [11]. Moreover, VHL-based HDAC6 degraders were capable of degrading HDAC6 without causing IKZFs degradation. HDAC6 seems to be very sensitive to PROTAC-mediated degradation, probably because of its dominant cytoplasmic localization, whereas most other isoforms are localized in the nucleus [139]. The degradation of class I HDACs was achieved by recruiting benzamide-based HDACi CI-994 as the warhead [140]. The degradation effect of this PROTAC was more significant on HDAC1-2 in comparison to HDAC3, which can probably be attributed to the HDAC binding preference of the warhead.

MDM2 is a known E3 ubiquitin ligase that acts as a powerful inhibitor of the tumor suppressor P53 [141]. Overexpression of MDM2 has been often observed in various human cancers and can contribute to tumorigenesis [142, 143]. Therefore, activation of the p53 pathway by antagonizing MDM2 attracts great interest for developing targeted tumor therapeutics. Several MDM2 SMIs including nutlins (nutlin 1, nutlin 2, nutlin 3) and their derivatives have been developed and exhibit good activities against various solid and hematologic malignancies, but can also cause hematopoietic suppression and gastrointestinal toxicity [144]. In addition, MDM2 inhibitor can lead to a significant compensatory increase in MDM2 expression, which can lead rapid degradation of p53 upon clearance of the MDM2 [46]. In 2018, Li et al. employed the PROTAC strategy to design small-molecule degraders of MDM2, including MD-224 as a lead degrader that targets MDM2 to CRBN for degradation [46]. MD-224 potently induced rapid degradation of MDM2 in leukemia cells. Moreover, MD-224 is capable of achieving complete and durable tumor regression in a RS4-11 xenograft model by intravenous administration. As expected, MD-224 can only inhibit the growth of leukemia cells carrying the wild-type P53 but not those with P53 mutants [46].

Polo-like kinase 1 (PLK1) is a Ser/Thr protein kinase which belongs to the CDC5/Polo subfamily. It is highly expressed in many different types of tumors, including hematologic malignancies such as AML and thus is an attractive tumor therapeutic targets. Mu et al. reported a CRBN-based BRD4/PLK1 dual degrader, HBL-4, which is more potent than its parental inhibitor BI2536 [145]. Interestingly, in an earlier report by the same group, a BI2536 derivative was used to build a PROTAC (compound 22f), which also displayed potent anti-proliferative activity in RS4-11 leukemia cells [146]. In that study, compound 22f only degraded BRD4. Therefore, it would be of great interest to test if compound 22f can also induce PLK1 degradation.

Polycomb repressive complex 2 (PRC2) is one of the two major classes of polycomb proteins, which have been broadly linked to hematologic malignancies [147‐149]. The PRC2 is composed of an enhancer of zeste homolog 2 (EZH2), EED, SUZ12, RBAP46/48, and AEBP2 [150]. EZH2 acts as the core catalytic subunit of PRC2, and EED and SUZ12 subunits are indispensable for EZH2 catalytic activity [151]. As PRC2 is highly associated with hematopoietic differentiation and proliferation [152], EZH2, EED, or SUZ12 loss-of-function mutations are related to the increased activity of hematologic stem cells (HSCs) and progenitor cells [153]. High frequency of mutations in EZH2, EED, and SUZ12 are observed in patients with early T cell precursor ALL (ETP-ALL) [154, 155]. Effective inhibition of PRC2 catalytic activity has been achieved by targeting both EED and EZH2. The discovery of deazaneplanocin A (DZNep) is the first attempt to inhibit EZH2 function, which could effectively deplete EZH2, SUZ12, and EED at cellular levels [156]. Later, a series of EZH2 inhibitors were developed for targeting wild-type and/or mutated EZH2 [157]. Due to the critical role of EED in EZH2 activity, the effect of EZH2 inhibitors could be phenocopied by SMIs targeting the EED [158, 159]. Although EZH2 inhibitors are promising antitumor agents, preclinical data suggest that drug resistance could be generated by secondary mutations in both wild-type and mutant EZH2 alleles [160].

In 2020, two separate research groups successfully converted the EED inhibitors into PROTACs [47, 161]. Interestingly, both groups used the VHL ligand to design the PROTACs, but used different EED inhibitors as the warhead, as well as different strategies for linker optimization. The most potent compound named PROTAC2 developed by Bloecher’s group selectively degraded EED, EZH2, and SUZ12, and reduced the proliferation of EZH-dependent tumor cells Karpas422 [161]. The most potent compounds named UNC6852 developed by James’ group also selectively degraded EED, EZH2, and SUZ12, and potently inhibited the proliferation of DLBCL cell lines with a wild-type and Y641N mutant EZH2 [47]. Although the EED-based PROTACs could degrade EZH2, so far no selective PROTACs based on EZH2 inhibitors were reported, until 2019 when Jian et al. developed MS1943, an EZH2 inhibitor-based degrader, via hydrophobic tagging strategy, which was designed by linking an EZH2 inhibitor to a bulky adamantyl group [162]. MS1943 could effectively degrade EZH2 and selectively kill EZH2-dependent TNBC cells, whereas EZH2 inhibitors are ineffective in blocking the proliferation of TNBC cells.

ATP-dependent chromatin remodeling complexes (CRCs) or remodelers use the energy of ATP hydrolysis to change the packaging state of chromatin and thus regulate cellular processes related to chromatin structure such as transcription, DNA repair, DNA replication, and decatenation of chromosomes during mitosis [48, 163]. Among the four distinct families of CRCs (SWI/SNF, ISWI, CHD, INO80), SWI/SNF complexes are the most implicated in diseases with more than 20% of human cancers bearing mutations in the genes encoding mammalian SWI/SNF (mSWI/SNF) subunits [163]. mSWI/SNF complexes are separated into two forms: BRG-/BRM-associated factor (BAF) and polybromo-associated BAF (PBAF). BAF and PBAF complexes share numerous subunits, such as BRG1 (SMARCA4) and BRM (SMARCA2) ATPases, which differ by incorporating different peripheral subunits. Mutations in BAF and PBAF subunits appear to drive different malignancies [163, 164]. SMARCA4 and SMARCA2 are ATPases that are mutually exclusively present in BAF and PBAF complexes. A study found that leukemia cells rely on SMARCA4 to support their oncogenic transcriptional program [165]. Whereas SMARCA2 is essential for the growth of tumor cells that harbor loss of function mutations in SMARCA4 [166]. These findings have inspired great interest in suppressing the bromodomain of SMARCA2 and SMARCA4 through SMIs [166‐168]. However, such inhibitors have failed to inhibit tumor cell growth [48] because the ATPase domain but not the bromodomain is essential for tumor cell growth [169]. By converting a non-functional bromodomain ligand to PROTAC molecules, Farnaby et al. successfully developed SMARCA2/4 degraders, including ACBI1. ACBI1 potently induced SMARCA2/4 and PBRM1 degradation in the AML cell line, and inhibited cell growth in multiple cancer cell lines [48]. Degradation of SMARCA2/4 and PBRM1 subunits may disrupt the integrity of the BAF and PBAF complexes, resulting in the dissociation of other subunits. Meanwhile, SMIs that target the allosteric binding site of the SMARCA2/4 ATPase domain and are orally bioavailable with in vivo antitumor activity have been reported [170]. However, due to the dual inhibition of SMARCA2 and SMARCA4, these allosteric inhibitors have a dose-limiting tolerability issue that hampered their further exploration. It would be interesting to see if converting these dual SMARCA2 and SMARCA4 inhibitors to PROTACs would result in the selective degradation of SMARCA2 or SMARCA4.

Signal transducer and activator of transcription 3 (STAT3) belongs to the STAT family of proteins, which are both signal transducers and TFs [171]. Aberrant activation of STAT3 is associated with many human malignancies [172‐174], because STAT3 regulates numerous genes involved in tumorigenesis, progression, and drug resistance, and thus it has become an attractive cancer target [172]. Several SMIs that block the dimerization of STAT3 have reached the clinical development stage, but only exhibited very limited clinical activity [173]. One major reason for their ineffectiveness is that monomeric STAT3 also has transcriptional activity [175]. To overcome the shortcoming, Wang’s group developed a series of STAT3 PROTACs, including SD-36 [3]. SD-36 efficiently degraded STAT3 protein in MOLM-16 AML cells and several anaplastic ALCL cell lines (SU-DHL-1, DEL, and KIJK). Interestingly, although SD-36 also binds to other STAT proteins, such as STAT1 and STAT4, it only degrades STAT3. In addition, SD-36 effectively degraded mutated STAT3 proteins in tumor cells. Further, SD-36 had desirable PK/pharmacodynamic (PD) properties, was well tolerated and achieved complete and long-lasting tumor regression in mice xenografted with AML and ALCL cell lines. More examples of STAT3 degraders were found in another study from Wang’s lab [176]. This study provides an excellent example that PROTAC technology can be used to target a previously undruggable target such as STAT3. However, STAT3 PROTACs may have some limitations in targeting leukemia than lymphoma, because SD-36 demonstrates potent activity in five of nine lymphoma cell lines, while showing activity in only one of nine AML cell lines [3]. Developing PROTACs that can target other STAT family proteins may add an additional layer of benefits in the treatment of hematologic malignancies.

In addition to the above targets, there are also some less studied PROTAC targets (SIRT2, PCAF/GCN5, BRD7, BRD9, IRAK4, TRIM24, and RIPK2) that are implicated in various malignancies, including leukemia and lymphoma. Schiedel et al. reported a PROTAC against sirtuin 2 (SIRT2). This PROTAC selectively induced SIRT2 degradation in HeLa cells and led to more pronounced acetylation of the microtubule network than the corresponding inhibitor [177]. P300/CBP-associated factor (PCAF) and general control nonderepressible 5 (GCN5) are multidomain epigenetic proteins, containing an acetyltransferase and a bromodomain, which are involved in many cellular processes including proliferation, DNA damage repair, and inflammation [178]. A PCAF/GCN5 PROTAC, named GSK983, efficiently degraded PCAF and GCN5 in THP1 acute monocytic leukemia cells [179]. BRD7 and BRD9 are bromodomain-containing subunits of the BAF and PBAF complexes, respectively [164]. Brander’s group reported a selective degrader of BRD9, named dBRD9, which efficiently degraded BRD9 and exerted a 10–100-fold increase in antiproliferative activity than the corresponding inhibitor against MOLM-13 and other AML cell lines [180]. Subsequently, Ciulli’s group reported a VHL-based dual BRD7 and BRD9 degrader, VZ185, which was found to be more effective than the corresponding inhibitor in reducing the viability of EOL-1 acute myeloid eosinophilic leukemia cells and A-204 malignant rhabdoid tumor cells [181]. The hyperactivation of interleukin-1 receptor-associated kinase 4 (IRAK4), a key mediator of innate immunity, is linked with autoimmune diseases and cancer [182]. IRAK4 has both enzymatic (kinase-dependent) and scaffolding (kinase-independent) functions [183]. Nunes et al. reported a VHL-based PROTAC that was capable of degrading IRAK4 in peripheral blood mononuclear cells and dermal fibroblasts with DC₅₀ values of 151 nM and 36 nM, respectively [184]. In addition, PROTACs targeting tripartite motif-containing protein 24 (TRIM24), receptor-interacting protein kinase 2 (RIPK2), and the Janus kinase family enzymes for degradation may also have the potential to be used to treat certain types of hematologic malignancies but require further studies [185‐187].

Calreticulin gene (CALR) mutations frequently occur in Philadelphia chromosome (Ph)-negative myeloproliferative neoplasms (MPNs), including essential thrombocythemia (ET), primary myelofibrosis (PMF), and polycythemia vera (PV) [188‐190]. CALR mutations are mutually exclusive with JAK2 or MPL mutations in MPNs, suggesting common oncogenic pathways. CALR mutations are located in exon 9 and generate a + 1 base-pair translation frameshift. Although the extent of the C-terminal alterations vary among CALR mutations, all resulting variants share a loss of C-terminal 27 amino acids with a concomitant gain of a novel peptide consisting of 36 amino acids.

CALR mutations has been shown to be pathogenic and serve as a disease-initiating event to favor the expansion of the megakaryocytic lineage by specific activation of the thrombopoietin receptor and uncontrolled activation of the JAK2/STAT pathway [191, 192]. The oncogenic CALR mutant protein could be an excellent PROTAC target to treat MPN patients with CALR mutations. It would be ideal to design the PROTAC against all CALR mutant proteins using specific ligand for the unique 36 amino acids in the C-terminal of all CALR mutants and a megakaryocytic lineage specific E3 ligase.

Nucleophosmin (NPM1) is one of the most commonly mutated genes in AML, occurring in approximately 50% of adult and 20% of childhood AML with normal karyotypes [193, 194]. Due to its clinical significance, NPM1-mutated AML was categorized as a distinct leukemic entity in the WHO-2016 classification [195]. NPM1 mutations are heterozygous and restricted to exon-12, causing a frameshift to eliminate the nucleolar localization sequence of NPM1 and to generate a new C-terminal nuclear export signal motif. Therefore, mutations in NPM1 result in mislocalization of NPM1 mutant proteins to the cytoplasm (NPM1c⁺). NPM1c⁺ acts as a dominant negative mutant for NPM1 function and has a gain-of-function role in pathogenesis of the AML phenotype [196, 197]. Although there are many different strategies in targeting NPM1 such as inhibition of translocation of NPM1c⁺ [198] and proteasomal degradation of NPM1c⁺ [199], the unmet needs to develop effective drugs targeting NPM1 is still growing. The development of a PROTAC specifically targeting NPM1c⁺ could be a superb strategy. Giving the distinct cellular localization of WT NPM1 and NPM1c⁺, a cytosolic E3 ligase would effectively eliminate the degradation of WT NPM1, if the ligand is not specific enough to NPM1c⁺. In addition, efforts could be directed to screen specific ligands for type A NPM1 mutation, which accounts for ~ 80% of the NPM1 mutations and has a unique 12 amino acid peptide in the C-terminal including a lysine.

Chromosomal translocations are very common in hematologic malignancies [200, 201]. It is well established that chromosomal translocations promote the development of hematologic malignancies either through the formation of oncogenic fusion protein or through oncogene activation by a stronger promoter or enhancer [200, 201]. Both the “alien” oncofusion proteins and the miss-activated oncoproteins are potential PROTAC targets. Indeed, several misactivated oncoproteins such as Bcl-2 and Bcl-6, and oncofusion proteins such as BCR-ABL have been targeted by PROTACs for specific subtypes of hematologic malignancies (see above sections in this review). Theoretically, these oncofusion proteins could be ideal PROTAC targets, since they are “alien” to human cells and the side effects would be easier to manage. The most challenging step for such PROTAC development will be the screening for relative specific ligands for such oncofusion proteins. The available crystal structures for some of these oncofusion proteins will aid this endeavor. The oncofusion proteins are often specific for subtype of hematologic malignancies, which also have potential value for PROTAC development in terms of E3 ligase selections, e.g., there are four most prevalent oncofusion proteins in AML: PML-RARα, AML1-ETO, CBFβ-MYH11, and MLL-fusions, while ETV6-RUNX1 and TCF3-PBX1 fusions are common in B-cell ALL [202].

ARV-110 is an orally bioavailable AR PROTAC that entered into a clinical trial to treat patients with metastatic castrate-resistant prostate cancer (NCT03888612) in 2019. Recently, ARV-110 was reported to have an acceptable safety profile and appears clinically effective in a patient with metastatic castration-resistant prostate cancer (https://​ir.​arvinas.​com/​node/​7296/​pdf). ARV-471 is a PROTAC designed to specifically target ER. In preclinical studies, ARV-471 demonstrated near-complete ER degradation in tumor cells, and induced robust tumor shrinkage when dosed as a single agent in multiple ER-driven xenograft models. ARV-471 is currently in a phase I clinical trial (NCT04072952). Its safety, tolerability, pharmacokinetics, and anti-tumor activity in patients with ER⁺/HER2⁻ locally advanced or metastatic breast cancer (https://​ir.​arvinas.​com/​node/​7296/​pdf) will likely be reported in the coming months. A positive result from these PROTAC clinical studies will provide proof-of-principle that PROTACs can be developed as effective antitumor therapeutics.

Although PROTACs have been developed to successfully target a number of targets associated with hematologic malignancies, none has entered into the clinical trials. However, several PROTACs have shown promising preclinical results in the treatment of hematologic malignancies [3, 4, 44], and will likely be evaluated in the clinic in the near future. In addition, many pharmaceutical companies have invested heavily in the development of their antitumor PROTAC pipelines. For example, Nurix is going to file an IND application with the FDA at the end of this year to test its BTK PROTAC in patients with B cell malignancies (https://​www.​nurixtx.​com/​research-development/​pipeline/​), Kymera is advancing a STAT3 PROTAC as one of their drug candidates in oncology (https://​www.​kymeratx.​com/​pipeline/​), and Dialectic Therapeutics (DT) is pursuing preclinical development of the Bcl-xL PROTAC DT2216 for the treatment of hematologic malignancies and solid cancer tumors in 2021 (https://​www.​dtsciences.​com/​#science). Therefore, we expect to witness more clinical studies on PROTACs for hematologic malignancies in the near future.