Selective Elimination of Senescent Cancer Cells by Galacto-Modified PROTACs

Abstract

Although selective and effective clearance of senescent cancer cells can improve cancer treatment, their development is confronted by many challenges. As part of efforts designed to overcome these problems, prodrugs, whose design is based on senescence-associated β-galactosidase (SA-β-gal), have been developed to selectively eliminate senescent cells. However, chemotherapies relying on targeted molecular inhibitors as senolytic drugs can induce drug resistance. In the current investigation, we devised a new strategy for selective degradation of target proteins in senescent cancer cells that utilizes a prodrug comprised of the SA-β-gal substrate galactose (galacto) and the proteolysis targeting chimeras (PROTACs) as senolytic agents. Prodrugs Gal-ARV-771 and Gal-MS99 were found to display higher senolytic indexes than those of ARV-771 and MS99. Significantly, results of in vivo studies utilizing a human lung A549 xenograft mouse model demonstrated that concomitant treatment with etoposide and Gal-ARV-771 leads to significant inhibition of tumor growth without eliciting significant toxicity.

Graphical Abstract

INTRODUCTION

Cellular senescence is a process in which cells remain viable and metabolically active, but are incapable of undergoing additional cycles of division.¹ In the process, the pro-inflammatory senescence-associated secretory phenotype (SASP) factors are produced.^2–3 These factors can promote neoplastic growth, therapy resistance, immunosuppression, metastasis and angiogenesis.^4–9 Moreover, senescent cancer cells can remain dormant and viable for long periods of time, which poses a risk for tumor relapse.¹⁰ Therefore, to improve long-term outcomes from cancer treatment, manipulation and elimination of senescent cells are critically important, but present a grand challenge.¹¹

Senolytic drugs that selectively destroy senescent cells have emerged as viable modes for treatment of disorders related to senescence.^12–14 Senolytic agents, initially developed because of their ability to transiently switch off senescence-associated antiapoptotic pathways,^15–17 have been derived for the most part from anticancer drugs such as inhibitors of Bcl-2 family members,^{15, 18–19} HSP90,²⁰ and MDM2.^21–22 Despite the encouraging results that have come from clinical studies,²³ these efforts have shown that senolytics suffer from toxicity issues owing to off-target effects associated with non-significant differences in the expression levels of their targets between senescent cells and normal cells.²⁴

In the recent past, proteolysis-targeting chimeras (PROTACs) have surged as a new therapeutic modality for the treatment various diseases including cancer.^25–26 Distinct from traditional small molecule inhibitors that bind to malfunctioning proteins to block their functions, PROTACs instead degrade their protein targets by exploiting the cellular ubiquitin-proteasome system.^25–26 A typical PROTAC molecule contains a ligand that recognizes the target protein linked to another ligand that recruits a specific E3 ubiquitin ligase. One example of this approach is found in studies by Zheng and colleagues in which the PROTAC molecule PZ15227 that derives from the BCL-2/BCL-XL inhibitor ABT263 was designed and evaluated as a senolytic agent.²⁷ Notably, this PROTAC displays lower toxicity to platelets than ABT263, but an equal or slightly higher activity against SnCs. Moreover, PZ15227 effectively clears senescent cells in naturally aged mice without inducing severe thrombocytopenia. However, like other PROTACs, this agent does not avoid systemic toxicity arising from off-target degradation and disruption of regular cellular proteostasis.^28–29 Nonetheless, the study by Zheng et al. suggests that the emerging PROTAC technology holds great potential as a viable strategy for effective elimination of senescent cells.

Senescence-associated β-galactosidase (SA-β-gal) is a broadly exploited biomarker for senescent cells as the enhanced activity of this lysosomal enzyme is a common feature of these types of cells.³⁰ Therefore, the biomarker has been generally used for probing senescent cells,^31–36 as well as the platform in prodrug design.^37–45 A basic feature of these senescent cell-targeting prodrugs is that they contain β-galactose moieties linked to known drugs or targeted small molecule inhibitors. For example, Tuo, Deng and colleagues designed a gemcitabine prodrug that is selectively de-caged by the action of SA-β-gal to induce apoptosis in senescent cells.³⁸ Munoz-Espin et al. described a galactose-encapsulated prodrug that undergoes selective SA-β-gal induced liberation of navitoclax in senescent cancer cells.⁴⁰ Similarly, a study by Li and coworkers demonstrated that the acetylated β-galactose containing prodrug Gal-LIN is activated by SA-β-gal to release the neddylation inhibitor linifanib (LIN) in senescent cancer cells.⁴⁴ However, it is known that frequent treatment of cancer cells with the chemotherapeutic components of these prodrugs can promote drug resistance.²⁴ Moreover, high doses of the prodrug conjugates are required to fully suppress a single senescent cell antiapoptotic pathway (SCAP).⁴⁵ This requirement creates a particular challenge related to therapeutic selectivity because a significant percentage of senescent cells are located in normal tissues. These conjugates can disrupt the structural integrities of tissue or vascular endothelial cells, leading to fibrosis and the collapse of liver and perivascular tissue.^46–47 These issues significantly limit the use of these agents especially in elderly individuals or those with co-morbidities. This underscores the urgent need for the development of a new family of cancer-selective senolytics.

The aforementioned challenges associated with existing senolytic strategies stimulated us to design a new family of SA-β-gal activatable PROTAC prodrugs (Gal-PROTACs) to improve senescent cell selectivity by promoting the specific degradation of proteins of interest (POI) in malignant senescent cells. Notably, our newly designed PROTAC prodrugs would provide the unique catalytic ability to selectively prompt protein degradation in a sub-stoichiometric manner. This would eliminate the requirement of high dosages, thus alleviating toxicities and drug resistance associated with inhibitor based senolytics.^48–51 To assess this strategy, we designed and prepared Gal-PROTACs, and investigated their ability to selectively ablate cancer senescent cells. In this study, we found that Gal-ARV-771 and Gal-MS99 prodrugs, which are designed based on ARV-771²⁹ and MS99⁵² PROTACs, are selectively activated by SA-β-gal in etoposide-induced senescent cancer cells. Both agents subsequently catalyze ubiquitin-proteasome-mediated degradation of their respective target proteins, BRD4 and NPM-ALK and effectively induce apoptosis in senescent cells. In addition, we determined that Gal-ARV-771 and Gal-MS99 have higher senolytic indexes than the respective parent compounds ARV-771 and MS99. Furthermore, in vivo studies show that concomitant treatment of human lung A549 xenografts with etoposide and Gal-ARV-771 leads to significant inhibition of tumor growth without any observable toxicity.

RESULTS AND DISCUSSION

Design, synthesis and trigger release studies of Gal-ARV-771.

We based our fundamental design strategy on the premise that incorporation of the SA-β-gal sensitive galactose (Gal) moiety into PROTACs would form Gal-PROTAC prodrugs, in which proper caging of a key moiety in the PROTAC would abolish its degradation activity. In addition, we postulated that proper placement of the Gal group into the E3 ubiquitin ligase VHL ligand would give the Gal-PROTAC general applicability.⁵³ It is known that the hydroxyl group in the VHL ligand plays a critical role in recruitment of VHL E3 ubiquitin ligase^54–55 and that caging the −OH group with folate,⁵⁶ aptamers⁵⁷ and a bioorthogonal TCO group^58–60 blocks its degradation activity. These observations guided us to design a Gal caged PROTAC, in which the Gal moiety is connected via a carbonate linked 4-hydroxyl-benzyl alcohol group to the hydroxyl group of the VHL-based bromodomain (BRD) degrader, ARV-771²⁹ (Figure 1A). The results of docking studies provided support for the expectation that the generated PROTAC, Gal-ARV-771, would possess decreased interactions between the VHL ligand and residues in the VHL protein binding site (Figure S1). It is expected that the Gal-ARV-771 would be inactive in media and that it would be selectively activated by SA-β-gal and an esterase present in senescent cells to release the protein degrader ARV-771.

Figure 1.

A) The pathway for senescence-associated β-galactosidase (SA-β-gal) triggered release of PROTACs from the β-galactose caged PROTAC prodrug Gal-ARV-771. B) The reaction process of porcine liver esterase (E) and SA-β-gal co-catalyzed hydrolysis of Gal-ARV-771 (100 μM, pH 7.4 phosphate buffer at 37 °C for 16 h) monitored by UFLC.

To establish proof of concept for our prodrug design strategy, Gal-ARV-771 was synthesized and characterized (Scheme S1), and then tested in biological studies. First, we examined the stability of Gal-ARV-771 by incubating it (100 μM) in 10% Fetal bovine serum (FBS) containing DMEM (Dulbecco’s Modified Eagle’s Medium). Real-time UFLC (Ultra-Fast Liquid Chromatography) monitoring revealed that Gal-ARV-771 is stable over a 24 h period (Figure S2). In contrast, in the presence of both esterase [porcine liver (catalog no. 9016-18-6), hydrolysis of the acyl group] and the human lysosomal β-galactosidase GLB1,^{38, 61} Gal-ARV-771 is nearly completely consumed within 16 h to cleanly form ARV-771 (Figure 1B). Notably, in the absence of the esterase, GLB1 does not promote hydrolysis of the glycosidic bond in Gal-ARV-771 because the acylated Gal moiety in the prodrug is not recognized by GLB1. These studies validate our working hypothesis that Gal-ARV-771 releases ARV-771 through the combined action of an esterase and β-galactosidase.

Gal-ARV-771 triggers ubiquitin-proteasome-dependent degradation of BRD4 in senescent A549 cells.

The PROTAC BET degrader ARV-771, developed previously for degradation of BRD4 protein in cancer cells,²⁹ was employed to demonstrate the efficacy of Gal-ARV-771 as a prodrug for inducing selective degradation of BRD4 protein in senescent cells. Therefore, senescent A549 cells was produced by low-dose treatment with etoposide.⁶² Quantification of the SA-β-gal activity of the widely employed biomarker for senescent cells revealed enhanced expression (Figure 2A, 2B and S3B).³⁰ In a control experiment, we observed that treatment of both normal A549 (n-A549) cells and etoposide-induced senescent A549 (s-A549) cells with ARV-771 (50 and 100 nM) induced degradation of BRD4 (Figure 2C). However, Gal-ARV-771 (100 nM) displayed degradation activity only in the s-A549 cells (Figure 2D and Figure S6B), indicating that this prodrug is cell permeable and converted to ARV-771 by a combination of esterase and SA-β-gal selectively expressed in the senescent cells. It is believed that Gal-ARV-771 enters the cells via endocytosis. HeLa (Figure S4) and U87 (Figure S5) cancer cells lines, and their respective etoposide induced senescent counterparts (Figure S3A) were utilized to demonstrate that the approach of inducing BRD4 degradation using a prodrug may be applicable to multiple tumor types. Discriminative degradation of BRD4 only took place in senescent cells. This is in accord with the outcomes observed in the A549 cells. Experiments were performed with the proteasome inhibitor MG-132, free VHL ligand, or NEDD8-activating enzyme (NAE) inhibitor MLN-4924 to confirm that BRD4 degradation is ubiquitin proteasome dependent. The results demonstrate that MG-132, VHL ligand and MLN-4924 inhibit the degradation abilities of both Gal-ARV-771 and ARV-771 (Figure 2E and Figure S6A), showing that Gal-ARV-771 and ARV-771 both depend on the ubiquitin-proteasome pathway.

Figure 2.

Degradation ability of Gal-ARV-771 in senescent A549 cells. SA-β-gal staining: control (A) and etoposide (5 μM) induced senescent A549 cells (B). C) Western blot detection of BRD4 in n-A549 and s-A549 cells treated with ARV-771 for 16 h. D) Western blot analysis of BRD4 in n-A549 and s-A549 cells treated with ARV-771 or Gal-ARV-771 for 16 h. Gal-ARV-771 had DC₅₀ = 18.3 nM in senescent A549 cells. E) Western blot analysis of BRD4 levels in s-A549 cells co-treated with 5 μM MG-132 or 10 μM VHL ligand for 16 h.

Gal-ARV-771 expresses selective activity against etoposide-induced senescent A549 cells.

Based on our results, Gal-ARV-771 has a protein degradation ability in s-A549 cells that is similar to that of ARV-771. Cell viability assays were performed to substantiate the expectation that Gal-ARV-771 is selectively activated to form ARV-771 only in senescent cells. Different doses of ARV-771 and Gal-ARV-771 were applied to s-A549 and n-A549 cells for 72 h. Cell viability assays determined that the IC₅₀ of Gal-ARV-771 was 3.29 μM for n-A549 and 640 nM for s-A549 cells (Figure 3A). Moreover, the senolytic index of Gal-ARV-771 between n-A549 and s-A549 cells is 5.17. In contrast, the IC₅₀ values for ARV-771 (Figure 3B) show that it is highly active against both n-A549 (354 nM) and s-A549 (603 nM) cells. In addition, the IC₅₀ of Gal-ARV-771 and ARV-771 in HeLa (Figure S4B, C) and U87 cells (Figure S5B, C) show that Gal-ARV-771 has a higher senolytic index than ARV-771.

Figure 3.

Activity of Gal-ARV-771 in senescent A549 and HeLa cells. A, B) Cell toxicity of n-A549 and s-A549 cells treated with ARV-771 or Gal-ARV-771 for 72 h. C, D) Flow cytometry profiles showing the effects of ARV-771 and Gal-ARV-771 on the induction of apoptosis in normal A549 and senescent A549 cells. Cells were cultured with 1 μM of ARV-771 or Gal-ARV-771 for 24 h, and then flow cytometry analysis was carried out with FITC and PI staining.

Furthermore, as a result of the dynamic and highly diversified nature of the senescence levels, other types of senolytic drugs show different apoptotic activities in various senescent cell types.^63–64 However, Gal-ARV-771 can induce cell independent apoptosis. To validate these findings, flow cytometry assays were carried out to determine the antitumor activity of Gal-ARV-771 in senescent cancer cells. In normal A549 cells (Figure 3C), compared to Gal-ARV-771 (8.07%), a higher percentage (20.1%) of early apoptosis was observed with ARV-771 (1.0 μM). However, in the senescent counterparts (Figure 3D), both ARV-771 and Gal-ARV-771 (1 μM) have similar anticancer activity (17.0% and 16.7%). The results presented here clearly demonstrate that Gal-ARV-771 selectively promotes apoptosis of senescent cancer cells.

Degradation of NPM-ALK protein by Gal-MS9 in senescent Karpas 299 cells.

To demonstrate the potential broad applications of our prodrug strategy for selective protein degradation in senescent cells, we used this platform to design and synthesize a second VHL-based PROTAC, Gal-MS99, directed against a totally independent protein target (Figure 4A). It was previously reported that MS99 degrades the nucleophosmin (NPM)–anaplastic lymphoma kinase (ALK) fusion protein.⁵² Because Karpas 299 cells overexpress NPM-ALK,⁶⁵ normal Karpas 299 (n-Karpas 299) and senescent Karpas 299 (s-Karpas 299) cells were employed to evaluate the degradation ability of Gal-MS99. The results demonstrate that Gal-MS99 does not promote degradation of NPM-ALK in n-Karpas 299 cells even at a concentration of 1 μM (Figure 4B), while in s-Karpas 299 cells this protein is degraded by both MS99 and Gal-MS99 (Figure 4C). Specifically, the IC₅₀ values for NPM-ALK degradation by Gal-MS99 in n-Karpas 299 and s-Karpas 299 cells are 2.162 μM and 454.8 nM, respectively (Figure 4D). The senolytic index between n-A549 and s-A549 cells is 4.75. However, the senolytic index of MS99 between n-A549 and s-A549 cells is only 1.59 (IC₅₀ = 368.8 nM in n-Karpas 299 and 232.4 nM in s-Karpas 299) (Figure 4E). Taken together, the results described here demonstrate that more effective treatment of tumor can be achieved by combined use of senescence-inducing therapy with senotherapy. Our data provide a preclinical proof-of-concept of the therapeutic values of Gal-ARV-771 as a potent PROTAC prodrug.

Figure 4.

NPM-ALK degradation properties of Gal-MS99 in senescent Karpas 299 cells. A, C) Western blot analysis of NPM-ALK protein in-Karpas 299 and s-Karpas 299 cells treated with MS99 and Gal-MS99 for 16 h. B, E) Cell viabilities of n-Karpas 299 and s-Karpas 299 cells treated with (B) Gal-MS99 and (E) MS99 for 72 h. D) Structure of Gal-MS99.

Gal-ARV-771 is well tolerated and has potent in vivo anti-senescent activity.

In the next phase of this investigation, we examined the tolerability and therapeutic benefit of Gal-ARV-771 utilizing the A549 xenograft mouse model. In this experiment, subcutaneous transplantation of A549 cells into the flanks of severe combined immunodeficient (SCID) mice. The tumor containing mice were treated with etoposide to induce senescence. Specifically, when tumors reached an average volume of 80–100 mm³, the mice were then treated with etoposide (5 mg/kg) and daily doses of Gal-ARV-771 (20 mg/kg) or ARV-771 (20 mg/kg) alone or in combination for 19 d (Figure 5A). Both ARV-771 and Gal-ARV-771 significantly inhibited tumor growth in etoposide treated mice (TGI% = 80% for ARV-771 and 74% for Gal-ARV-771) (Figure 5B). Notably, body weight measurements and daily observations showed that these agents were very well tolerated as no significant toxicity was observed (Figure 5C). In contrast, treatment of the mice with Gal-ARV-771 alone did not have a comparable effect on tumor growth inhibition, indicating that its therapeutic effect in the A549 mouse model requires induction of senescence by etoposide (Figure S7). Moreover, immunohistochemistry (IHC) assays revealed that treatment with etoposide and Gal-ARV-771 significantly reduced tumor expression levels of BRD4 (Figure 5D, E). Importantly, combination treatment also significantly reduced tumor cell proliferation (Ki-67 IHC) (Figure 5D, F), and concomitantly increased apoptosis (cleaved caspase-3) (Figure 5D, G). Moreover, compared to etoposide alone treatment, combination treatment reduced the expression level of senescent biomarker p21 (Figure S8). These findings suggest that the anti-tumor effect of the prodrug is caused by apoptosis of senescent cells. Taken together, the results described here demonstrate that more effective treatment of tumor can be achieved by combined use of senescence-inducing therapy with senotherapy. Our data provide a preclinical proof-of-concept of the therapeutic values of Gal-ARV-771 as a potent PROTAC prodrug.

Figure 5.

Significant inhibition of tumor growth by co-treatment with prodrug Gal-ARV-771 and etoposide observed in a human lung cancer xenograft mouse model. A) Schematic of regime for concomitant treatment on A549 xenograft-bearing mice. B) Tumor volumes and C) body weights of A549 xenografts in mice concomitantly treated with etoposide (5 mg/kg) and ARV-771 (20 mg/kg) or Gal-ARV-771 (Gal, 20 mg/kg). D) Representative histological staining of BRD4, Ki-67 and cleaved caspase 3 after final concomitant treatment. E-G) Quantification of BRD4, Ki-67 and cleaved caspase-3-positive cells in tumors from animals treated with vehicle, Gal-ARV-771, or etoposide and Gal-ARV-771 (n = 3 tumors per group). For quantification, analysis of a total of 3 fields per tumor was performed. *p < .05; **p < .01; ***p < .001; ****p < .0001.

Proteomic analysis demonstrates that Gal-ARV-771 selectively degrades BRD4 in s-A549 cells.

To evaluate the effects of Gal-ARV-771 on BRD4 in the senescent cells, we conducted global proteomic analyses of n-A549 and s-A549 cells. A multiplexed method was used to quantify protein levels in both normal and etoposide-induced senescent A549 cells treated with 50 nM Gal-ARV-771 (Figure 6A). Compared to that in n-A549 cells, the expression level of BRD4 was significantly lower in the s-A549 counterparts. We also identified changes in the senescent-related biomarker L1CAM,⁶⁶ which previous studies have shown to be induced in premature forms of chemically triggered senescence (Figure 6A).⁶⁷ L1CAM might represent a new potential senescence biomarker, especially since it is a cell membrane protein. The notable changes of the top five protein levels also included ARGAL and BTBD8 (Figure 6B). The information was valuable for the better understanding of the degradation mechanism of Gal-ARV-771 in s-A549 cells.. Taken together, the proteomics data further demonstrate the degradation ability of Gal-ARV-771 in chemically induced senescent cells.

Figure 6.

Proteomic analysis of Gal-ARV-771 in A549 cells. Normal A549 cells and etoposide-induced senescent A549 cells were treated 50 nM Gal-ARV-771 for 16 h. Lysates were subjected to proteomic analysis (Data represented three independent experiments). A) The distribution of all proteins (n = 4889) by volcano plot (relative protein abundance (log2FC) as a function of significance level (-log10P-Value)). B) Heat map representation of 6 individual sample abundances for top five up- and down-regulated proteins from proteomics data.

CONCLUSION

In summary, in this study we developed and assessed a unique and effective SA-β-Gal responsive prodrug strategy for selective degradation of senescent cancer cells. In a proof-of-principle investigation, we demonstrated that the PROTACs Gal-ARV-771 and Gal-MS99 are selectively activated in SA-β-Gal expressed cancer senescent cells to release protein degrading ARV-771 and MS99. Furthermore, these prodrugs decrease the cell viability of multiple types of senescent cancer cells suggesting broad therapeutic applicability. Importantly, we clearly showed that concomitant treatment with the promising senolytic prodrug Gal-ARV-771 and etoposide leads to effective inhibition of tumor growth in the A549 lung cancer mouse model. Perhaps the most significant feature of this investigation is that it demonstrates that Gal-PROTAC prodrugs have the unique ability to break down target proteins in a sub-stoichiometric and catalytic manner, which should minimize the development of drug resistance. Our collective findings define Gal-PROTAC prodrugs as a novel and powerful tool for the selective eradication of relapse-driving senescent cancer cells as well as mitigation of other human senescence-related disorders.

EXPERIMENTAL SECTION

General

Common materials or chemical reagents were purchased from commercial sources and used without further purification. The solvents were used by dry solvents system. All reactions were monitored by TLC or LC-MS. Purification was conducted on preparative flash column chromatography and preparative reversed-phase high performance liquid chromatography (RP-HPLC) with solvent systems specified. Nuclear magnetic resonance (NMR) spectra were recorded on automated Bruker AVIII-500 instruments. High-resolution mass spectra (HRMS) were recorded on a Bruker microTOF II instrument in positive ion mode using an Agilent G1969 API-TOF with an electrospray ionization (ESI) source. Ultraperformance liquid chromatography (UPLC) spectra for compounds were acquired using a Shimadzu LabSolutions system. All separations involved mobile phase of 0.1% TFA (v/v) in water (solvent A) and 0.1% TFA (v/v) in ACN (solvent B). The analytical HPLC conditions is linear gradient 5–95% (B %, v/v) at a flow rate of 1 mL/min over 30 min. The preparative HPLC conditions are linear gradient 5–95% (B %, v/v) at a flow rate of 17 mL/min over 40 min. The purity of end products Gal-ARV-771 and Gal-MS99 was determined by UPLC and their purities are > 95% (see SI).

Anti-BRD4 antibody (catalog no. 13440S) and anti-ALK antibody (catalog no. 3633T) were purchased from Cell Signaling Technology. Anti-GAPDH (catalog no. M7815), and Anti-Rabbit antibodies (catalog no. MFCD00162788) were purchased from Millipore Sigma. Anti-mouse lgG antibody (catalog no. A28177) was purchased from the Thermo Fisher. All primary antibodies were used at a 1:1000 dilution in 5% non-fat milk in Phosphate-buffered saline with 0.1% Tween-20 (PBST) buffer for western blot assay. All secondary antibodies were used at a 1:2000 dilution in 5% non-fat milk in Phosphate-buffered saline with 0.1% Tween-20 (PBST) buffer for western blot assay. β-Galactosidase from Escherichia coli (catalog no. 9031-11-2) and Esterase from porcine liver (catalog no. 9016-18-6) were purchased from Millipore Sigma.

Synthesis

4-O-(2,3,4,6-Tetra-O-acetyl-β-D-galactopyranosyl)-4-oxybenzaldehyde (1)⁶⁸

To the mixture of α-D-galactopyranosyl bromide (1.9 g, 4.63 mmol) and p-hydroxy benzaldehyde (622 mg, 5.09 mmol) in acetonitrile (49 mL) was added Ag₂O (2.14 g, 9.26 mmol) at room temperature and stirred for 16 h. After completion of the reaction, the reaction mixture was filtered through a celite pad and washed with ethyl acetate. The solvent was removed by rotavapor under reduced pressure, and the crude product was purified by silica gel flash chromatography, eluted with 30% EtOAc in hexane) to give an oil (1.6g, 76% yield). ¹H NMR (400 MHz, CDCl₃) δ 9.91 (s, 1H), 7.88 – 7.80 (m, 2H), 7.14 – 7.06 (m, 2H), 5.56 – 5.46 (m, 1H), 5.46 (s, 1H), 5.19 – 5.08 (m, 2H), 4.27 – 4.14 (m, 2H), 4.14 – 4.08 (m, 1H), 2.17 – 1.96 (m, 12H); ¹³C NMR (101 MHz, CDCl₃) δ 190.8, 170.4, 170.3, 170.2, 169.4, 161.4, 131.9, 116.8, 98.7, 71.4, 70.8, 68.5, 66.8, 61.4, 20.8, 20.8, 20.7, 20.7; LCMS (ESI, M + H⁺) calcd for C₂₁H₂₇O₁₁: 455, found 455.

4-(Hydroxymethyl)phenyl-(2,3,4,6-tetra-O-acetyl-β-D-galactopyranoside) (2)⁶⁸

To the solution of compound 1 (500 mg, 1.1 mmol) in CHCl₃/i-PrOH (33 mL, 3:1) in ice-bath condition, was added NaBH₄ (88 mg, 2.32 mmol) in two portions. The reaction mixture was brought to room temperature and stirred for 3 h, washed with 10% citric acid solution, saturated NaHCO₃, water, and brine. The mixture was filtered and dried over MgSO₄ and concentrated. Silica gel flash chromatography eluted with 50% EtOAc in hexane gave a white solid (313 mg, 63% yield). ¹H NMR (500 MHz, CDCl₃) δ 7.57 – 7.48 (m, 2H), 7.26 – 7.20 (m, 2H), 5.76 – 5.67 (m, 2H), 5.35 (dd, J = 10.4, 3.4 Hz, 1H), 5.27 (d, J = 7.9 Hz, 1H), 4.88 (s, 2H), 4.47 (dd, J = 11.3, 6.9 Hz, 1H), 4.40 (dd, J = 11.3, 6.4 Hz, 1H), 4.29 (td, J = 6.6, 1.2 Hz, 1H), 2.42 (s, 3H), 2.30 (d, J = 2.8 Hz, 6H), 2.25 (s, 3H); ¹³C NMR (126 MHz, CDCl₃) δ 170.4, 170.3, 170.2, 169.4, 156.5, 135.9, 128.5, 117.1, 99.8, 71.0, 70.8, 68.7, 66.9, 64.8, 61.4, 20.7, 20.7, 20.6; LC-MS (ESI, M + H⁺) calcd for C₂₁H₂₇O₁₁: 455, found 455.

4-(4-Nitro-phenoxycarbonyloxymethyl)phenyl]-(2,3,4,6-tetra-O-acetyl-β-D-galactopyranoside) (4)

The mixture of compound 2 (193 mg, 0.424 mmol), p-nitro-phenyl chloroformate (257 mg, 1.27 mmol), and pyridine (103 μL) in DCM (10 mL) was stirred at room temperature under N₂ gas. Completion of the reaction was checked by TLC. After completion of the reaction, it was washed with water and extracted with ethyl acetate. The organic portion was dried and concentrated and purified by silica gel flash chromatography (30–50 % EtOAc in hexane) as a white solid (199 mg, 76% yield). ¹H NMR (500 MHz, CDCl₃) δ 8.26 (d, J = 2.2 Hz, 1H), 8.25 (d, J = 2.2 Hz, 1H), 7.38 (d, J = 2.0 Hz, 1H), 7.37 – 7.35 (m, 2H), 7.34 (d, J = 2.2 Hz, 1H), 7.02 (d, J = 2.1 Hz, 1H), 7.01 (d, J = 2.1 Hz, 1H), 5.48 (dd, J = 10.5, 7.9 Hz, 1H), 5.45 (dd, J = 3.4, 1.1 Hz, 1H), 5.23 (s, 2H), 5.10 (dd, J = 10.4, 3.4 Hz, 1H), 5.05 (d, J = 7.9 Hz, 1H), 4.22 (dd, J = 11.3, 6.9 Hz, 1H), 4.14 (dd, J = 11.3, 6.4 Hz, 1H), 4.05 (td, J = 6.6, 1.2 Hz, 1H), 2.17 (s, 3H), 2.05 (d, J = 1.6 Hz, 6H), 2.00 (s, 3H); ¹³C NMR (101 MHz, CDCl₃) δ 170.5, 170.3, 170.2, 169.5, 157.5, 155.6, 152.5, 130.7, 129.2, 125.4, 121.8, 117.2, 99.5, 71.2, 70.8, 70.6, 68.6, 66.9, 61.4, 20.8, 20.8, 20.7. HRMS (ESI, M + Na⁺) calcd for C₂₈H₂₉NO₁₅Na: 642.1434, found 642.1429.

(2S,4R)-1-((S)-2-(5-(4-(4-((5-Chloro-4-((2-(Isopropylsulfonyl)phenyl)amino)pyrimidin-2-yl)amino)-5-isopropoxy-2-methylphenyl)piperidin-1-yl)-5-oxopentanamido)-3,3-dimethylbutanoyl)-4-hydroxy-N-((S)-1-(4-(4-methylthiazol-5-yl)phenyl)ethyl)pyrrolidine-2-carboxamide (MS99)

MS99 was synthesized following the literature procedure as an off white solid.^{69 1}H NMR (500 MHz, CD₃OD) δ 8.96 (s, 1H), 8.32 (d, J = 8.2 Hz, 1H), 8.20 (s, 1H), 7.98 (dd, J = 7.9, 1.6 Hz, 1H), 7.70 (t, J = 7.9 Hz, 1H), 7.50 (t, J = 7.7 Hz, 1H), 7.45 – 7.35 (m, 6H), 6.89 (s, 1H), 4.99 (p, J = 6.9 Hz, 1H), 4.72 (ddt, J = 13.2, 4.5, 2.1 Hz, 1H), 4.63 (dd, J = 6.2, 3.8 Hz, 2H), 4.57 (ddd, J = 11.4, 6.5, 2.7 Hz, 1H), 4.43 (dt, J = 4.4, 2.2 Hz, 1H), 4.11 (d, J = 13.3 Hz, 1H), 3.89 (d, J = 11.0 Hz, 1H), 3.75 (dd, J = 11.0, 3.9 Hz, 1H), 3.37 (p, J = 6.8 Hz, 1H), 3.23 (td, J = 13.1, 2.5 Hz, 1H), 3.04 (tt, J = 12.1, 3.4 Hz, 1H), 2.74 (td, J = 13.0, 2.7 Hz, 1H), 2.48 (s, 3H), 2.45 – 2.30 (m, 3H), 2.20 (s, 4H), 2.02 – 1.88 (m, 3H), 1.88 – 1.76 (m, 2H), 1.67 (ddd, J = 16.3, 8.4, 3.2 Hz, 1H), 1.57 (d, J = 7.2 Hz, 1H), 1.49 (t, J = 7.5 Hz, 3H), 1.31 – 1.19 (m, 12H), 1.06 (s, 9H). HRMS (ESI-MH+) calcd for C₅₆H₇₃ClN₉O₈S₂: 1098.4712, found 1098.4706.

(2R,3S,4S,5R,6S)-2-(Acetoxymethyl)-6-(4-((((((3R,5S)-1-((S)-2-(5-(4-(4-((5-chloro-4-((2-(isopropylsulfonyl)phenyl)amino)pyrimidin-2-yl)amino)-5-isopropoxy-2-methylphenyl)piperidin-1-yl)-5-oxopentanamido)-3,3-dimethylbutanoyl)-5-(((S)-1-(4-(4-methylthiazol-5-yl)phenyl)ethyl)carbamoyl)pyrrolidin-3-yl)oxy)carbonyl)oxy)methyl)phenoxy)tetrahydro-2H-pyran-3,4,5-triyl triacetate (Gal-MS99)

The solution of MS99⁶⁹ (5, 25 mg, 0.0227 mmol) and compound 4 (17 mg, 0.0273 mmol) in DCM (10 mL) was stirred at room temperature in the presence of DIPEA (8 μL) and catalytic DMAP for 48 h till completion. After completion, the solvent was dried and purified using prep HPLC to give an off-white solid (22 mg, 63%). ¹H NMR (500 MHz, CD₃OD) δ 8.85 (s, 1H), 8.42 (t, J = 8.7 Hz, 1H), 8.14 (d, J = 2.1 Hz, 1H), 7.91 (dt, J = 8.0, 2.0 Hz, 1H), 7.80 (d, J = 8.0 Hz, 1H), 7.65 (ddd, J = 8.6, 7.4, 1.6 Hz, 1H), 7.43 – 7.24 (m, 7H), 6.99 (t, J = 8.3 Hz, 2H), 6.81 – 6.72 (m, 1H), 5.42 (dd, J = 8.5, 3.5 Hz, 1H), 5.37 – 5.29 (m, 1H), 5.26 – 5.18 (m, 2H), 5.17 – 5.09 (m, 1H), 5.08 – 5.02 (m, 1H), 4.97 (dq, J = 9.9, 7.0 Hz, 1H), 4.68 (d, J = 13.3 Hz, 1H), 4.61 – 4.54 (m, 1H), 4.54 – 4.44 (m, 2H), 4.39 – 4.26 (m, 1H), 4.26 – 4.20 (m, 1H), 4.19 – 4.09 (m, 2H), 4.01 (d, J = 12.5 Hz, 1H), 3.85 (dt, J = 12.2, 4.4 Hz, 1H), 3.36 – 3.33 (m, 1H), 3.14 (dtd, J = 16.0, 13.2, 2.5 Hz, 1H), 3.00 – 2.89 (m, 1H), 2.74 – 2.64 (m, 1H), 2.45 (d, J = 2.0 Hz, 3H), 2.41 – 2.31 (m, 4H), 2.19 – 2.10 (m, 10H), 2.06 – 1.94 (m, 9H), 1.90 (q, J = 7.3 Hz, 2H), 1.84 – 1.70 (m, 2H), 1.46 (dd, J = 9.8, 7.0 Hz, 3H), 1.32 – 1.28 (m, 3H), 1.28 – 1.20 (m, 9H), 1.06 (s, 9H). ¹³C NMR (101 MHz, CDCl₃) δ 174.0, 171.9, 170.6, 170.4, 170.3, 169.6, 169.4, 160.5, 157.3, 157.2, 154.3, 152.4, 149.3, 144.6, 141.8, 136.2, 134.7, 131.7, 130.4, 129.8, 129.8, 128.3, 127.1, 126.7, 125.9, 125.7, 123.9, 117.1, 111.6, 105.7, 99.6, 76.2, 71.8, 71.2, 70.8, 69.7, 68.7, 66.9, 61.4, 58.9, 56.6, 53.9, 49.3, 38.4, 35.7, 35.2, 33.5, 33.0, 32.3, 29.7, 26.5, 22.1, 21.9, 20.8, 20.7, 20.7, 20.7, 18.8, 15.4, 14.0; HRMS (ESI-MH⁺) calcd for C₇₈H₉₇ClN₉O₂₀S₂: 1578.5979, found 1578.5972.

(2R,5R)-2-(Acetoxymethyl)-6-(4-((((((3R,5S)-1-((S)-2-(tert-butyl)-15-((S)-4-(4-chlorophenyl)-2,3,9-trimethyl-6H-thieno[3,2-f][1,2,4]triazolo[4,3-a][1,4]diazepin-6-yl)-4,14-dioxo-6,10-dioxa-3,13-diazapentadecanoyl)-5-(((S)-1-(4-(4-methylthiazol-5-yl)phenyl)ethyl)carbamoyl)pyrrolidin-3-yl)oxy)carbonyl)oxy)methyl)phenoxy)tetrahydro-2H-pyran-3,4,5-triyl triacetate (Gal-ARV-771)

The title compound was prepared in a similar manner to that of Gal-MS99 in 58% yield. ¹H NMR (500 MHz, DMSO-d₆) δ 8.99 (s, 1H), 8.46 (d, J = 7.6 Hz, 1H), 8.27 (t, J = 5.7 Hz, 1H), 7.48 (d, J = 8.7 Hz, 2H), 7.45 – 7.40 (m, 4H), 7.36 (dt, J = 7.9, 5.5 Hz, 5H), 7.02 – 6.95 (m, 2H), 5.48 (d, J = 7.8 Hz, 1H), 5.34 (d, J = 3.2 Hz, 1H), 5.28 (dd, J = 10.4, 3.5 Hz, 1H), 5.21 (dd, J = 10.3, 7.9 Hz, 1H), 5.17 (s, 1H), 5.09 (s, 2H), 4.92 – 4.87 (m, 1H), 4.55 – 4.49 (m, 1H), 4.44 (dt, J = 13.8, 8.4 Hz, 3H), 4.09 (d, J = 6.4 Hz, 2H), 3.99 – 3.93 (m, 1H), 3.91 (d, J = 2.9 Hz, 2H), 3.78 (dd, J = 11.9, 3.6 Hz, 1H), 3.53 (t, J = 6.5 Hz, 2H), 3.48 (t, J = 6.5 Hz, 2H), 3.42 (t, J = 5.8 Hz, 2H), 3.32 – 3.20 (m, 4H), 2.60 (s, 3H), 2.45 (s, 3H), 2.40 (s, 3H), 2.32 (dd, J = 13.8, 7.8 Hz, 1H), 2.14 (s, 3H), 2.03 (s, 3H), 2.00 (s, 4H), 1.94 (s, 3H), 1.78 (p, J = 5.5 Hz, 2H), 1.61 (s, 3H), 1.36 (d, J = 7.0 Hz, 3H), 0.94 (s, 9H). ¹³C NMR (126 MHz, DMSO-d₆) δ 184.9, 170.0, 169.8, 169.6, 169.2, 169.1, 168.7, 163.1, 158.4, 158.1, 156.6, 155.1, 153.7, 151.5, 149.9, 147.7, 144.6, 136.7, 135.3, 132.2, 131.1, 130.3, 130.2, 129.9, 129.7, 129.6, 128.9, 128.5, 126.3, 116.3, 97.6, 76.8, 70.4, 70.1, 69.4, 68.9, 68.7, 68.3, 68.0, 67.2, 67.0, 61.3, 58.0, 55.9, 53.8, 47.9, 40.1, 38.6, 37.4, 35.3, 34.6, 29.4, 26.1, 22.5, 20.5, 20.4, 20.4, 20.3, 15.9, 14.0, 12.7, 11.3.

Cell Culture

HeLa, A549, Karpas 299 and U87 cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% FBS (GIBCO, catalog no. 10437) and 1 % penicillin/streptomycin (Thermo Fisher, catalog no.30–002-Cl). The cells were grown at 37 °C with 5% CO₂.

Senescence-associated β-galactosidase (SA-β-Gal) staining

Briefly, A549 (HeLa or U87) cells were treated with 5 μM etoposide for 3 days. Then the cells were incubated with fresh medium for another day. After the treatment, the cells were washed with PBS twice, fixed with 4% paraformaldehyde for 15 min, and then incubated for overnight at 37 °C with the SA-β-Gal staining solution of the senescence β-galactosidase staining Kit (Cell signaling Technology, catalog no. 9860S) according to the manufacturer’s instructions. The cell images were taken by the ECHO Microscopes.

Senescence Cell Detection assay

The A549 cells were cultured in the 96-well plates with 3.0 × 10³ cells per well. After senescence incubation, firstly, the cell count normalization kit (Dojindo Molecular Technologies, catalog no. C544–02) was added to both normal A549 and senescent A549 cells for 30 min at 37 °C, 5% CO₂ incubator. The fluorescence was read by a Microplate Reader (Ex: 350 nm, Em: 461 nm). Then discard the supernatant and added lysis buffer and SPiDER–β-gal buffer from Senescence Cell Detection kit (Dojindo Molecular Technologies, catalog no. SG05–01) for 1 h. The fluorescence was read by a Microplate Reader (Ex: 510 nm, Em: 570 nm). Normalized SA-β-gal activity = (Fluorescent intensity of SA-β-gal) / (Fluorescent intensity of cell number). The graph was made by GraphPad software.

Cell viability assay

Tumor cells were plated in 96-well plates with 6.0 × 10³ cells in each well and subsequently incubated overnight in a moist atmosphere of 5% CO₂ and 37 °C. Then, different concentrations of test compounds or vehicles were added to triplicate wells. After incubation for an additional 72 h, 10 μL of CCK-8 solution (Dojindo Molecular Technologies, catalog no. CK04–11) was added to each well, then the plates were incubated for 1 – 4 h at 37 °C. The absorbance was read at 450 nm on a Microplate Reader. The values of IC₅₀ were calculated by the Logit method with the GraphPad software.

Immunoblot assay

Cells were lysed in RIPA buffer supplemented with protease inhibitors. The lysates (40–60 μg protein) were then resolved by 4%−12% Mini Protein Gel (Thermo Fisher) at 70 V for 10 mins and 200 V for 25 mins. Then the proteins were transferred from the gel to PVDF membrane (Bio-Rad) at 20 V for 120 mins. The membrane was incubated with primary antibody at 4 °C overnight, washed 3 times with PBST, incubated secondary antibody in 5% nonfat milk for 60 mins at room temperature and then washed 3 times with PBST. The membranes were detected under the SuperSignal West Atto Ultimate Sensitivity Chemiluminescent Substrate (catalog no. A38555).

Apoptosis assay

A549 cells were placed in 6-well transparent plates and then treated with test compounds and vehicle in a moist atmosphere of 5% CO₂ at 37 °C for 48 h. After that, the cells were washed twice with cold Cell staining Buffer (BioLegend, catalog no. 420201-BL) twice, and then were resuspended in Binding buffer (BioLegend, catalog no. 640914). 5 μL of FITC Annexin V (BioLegend, catalog no. 640914) and 10 μL of propidium Iodide Solution (BioLegend, catalog no. 640914) were added to the resuspended cell solution, which was then incubated for 15 min at room temperature in the dark. The analysis of stained cells was performed by a flow cytometer (BD FACS Canto II).

Mouse experiments

For establish subcutaneous tumor xenografts, 5 weeks ol NOD. Cg-Prkdc^scid/J females were injected subcutaneously with 2 × 10⁶ A549 cells in each flank. Tumors were measured with calipers every 2 days, and the tumor volume was calculated with the formula length × width²/2. When tumor volume reached an average of 80–100 mm³, mice were randomized and assigned to one of the control or therapy groups. Therapy was initiated with either vehicle, 5 mg/kg etoposide (eto, via intraperitoneal (i.p.) injection) three times per week, 20 mg/kg ARV-771 (via i.p. injection) 5 days ON/2 days OFF, 20 mg/kg Gal-ARV-771 (via i.p. injection) 5 days ON/2 days OFF, or a combination of the mentioned drugs. Mice were culled by cervical dislocation after 3 weeks of treatment.

IHC staining

Pre-loading. Place FFPE slides on the Leica Bond RXm racks and secure cover tile (Leica, catalog no. S21.4611) secured over the slide. Label slide with protocol for instrument to read.

Pre-Staining

1. Bake/Dewax

   1. Slides heated to 60°C for 30 minutes
   2. Heated until temperature reaches 72°C
   3. Dewax solution (Leica, catalog no. AR9222) dispensed and incubated for 30 seconds
2. HIER – Heat Induced Epitope Retrieval

   1. Bond Epitope Retrieval Solution 2 dispense (Leica, catalog no. AR9640)
   2. Bond Epitope Retrieval Solution 2 (EDTA based pH9) dispense (Leica, catalog no. AR9640)
   3. Bond Epitope Retrieval Solution 2 dispense (Leica, catalog no. AR9640) and incubated for 20 minutes
   4. Wash buffer (Leica, catalog no. AR9961) dispense, three times

Staining Protocol – the reagents in bold are part of Bond Polymer Refine Detection Kit DS9800 from Leica Biosystems.

(All dispenses are 150 μL unless otherwise noted)

1. Peroxide block for 5 minutes at room temperature

2. Wash buffer (Leica, catalog no. AR9590) dispense, three times

3. Primary antibody Leica Anti-BRD4 1:500 (Cell Signaling catalog, no. 13440S. Rb mAb), Anti-CC3 at 1:300 dilution (Cell Signaling, catalog no. 9661. Rb pAb), and Anti-KI67 at 1:900 dilution (Abcam, catalog no. ab15580. Rb pAb. 0.7mg/ml). Incubation for 15 minutes at room temperature

4. Wash buffer (Leica, catalog no. AR9590) dispense, three times

5. Polymer (anti-Rabbit Poly-HRP IgG) for 8 minutes at room temperature

6. Wash buffer (Leica, catalog no. AR9590) dispense for 2 minutes at room temperature, twice

7. Deionized water

8. Mixed DAB at room temperature

9. Mixed DAB for 10 minutes at room temperature

10. Deionized water, three times

11. Hematoxylin for 5 minutes at room temperature

12. Wash buffer (Leica, catalog no. AR9590) dispense

13. Deionized water dispenses

Post-Staining

1. Slides dehydrated in ethanol

   1. 80% ethanol – 2 minutes
   2. 100% ethanol – 2 minutes
   3. 100% ethanol – 2 minutes
   4. 100% ethanol – 2 minutes
2. Slides cleared in xylene

   1. Xylene – 3 minutes
   2. Xylene – 3 minutes
   3. Xylene – 3 minutes

3. 2 drops of mounting media (Leica, catalog no. 3801730)

4. Coverslip (Leica/Fisher, catalog no. 3800145ACS)

Proteomics assay

In-gel digestion.

Boiled clarified whole cell lysate (35 μg) was separated by 10% SDS-PAGE and stained with Bio-Safe Coomassie G-250 Stain (#1610786; Biorad, Hercules, CA). Each lane of the SDS-PAGE gel was cut into a single slice corresponding to the size BRD4 migrates on an SDS-PAGE gel. The gel slices were subjected to trypsin digestion and the resulting peptides were purified by C18-based desalting exactly as previously described.⁷⁰

Mass spectrometry and database search.

HPLC-ESI-MS/MS was performed in positive ion mode on a Thermo Scientific Orbitrap Fusion Lumos tribrid mass spectrometer fitted with an EASY-Spray Source (Thermo Scientific, San Jose, CA). NanoLC was performed as previously described.⁷¹ Tandem mass spectra were extracted from Xcalibur ‘RAW’ files and charge states were assigned using the ProteoWizard 2.1.x msConvert script using the default parameters. The fragment mass spectra were searched against the SwissProt_2022 database (Homo sapiens, 20402 entries) using Mascot (Matrix Science, London, UK; version 2.8.0.1) using the default probability cut-off score. The search variables that were used were: 10 ppm mass tolerance for precursor ion masses and 0.5 Da for product ion masses; digestion with trypsin; a maximum of two missed tryptic cleavages; variable modifications of oxidation of methionine and phosphorylation of serine, threonine, and tyrosine. Cross-correlation of Mascot search results with X! Tandem was accomplished with Scaffold (version Scaffold_5.1.2; Proteome Software, Portland, OR, USA). Probability assessment of peptide assignments and protein identifications were made using Scaffold. Only peptides with ≥ 95% probability were considered.

Label-free peptide/protein quantification and identification. Progenesis QI for proteomics software (version 2.4, Nonlinear Dynamics Ltd., Newcastle upon Tyne, UK) was used to perform ion-intensity based label-free quantification as previously described.^72–75 In brief, in an automated format, raw files were imported and converted into two-dimensional maps (y-axis = time, x-axis =m/z) followed by selection of a reference run for alignment purposes. An aggregate data set containing all peak information from all samples was created from the aligned runs, which was then further narrowed down by selecting only +2, +3, and +4 charged ions for further analysis. The samples were then grouped and a peak list of fragment ion spectra from only the top eight most intense precursors of a feature was exported in Mascot generic file (.mgf) format and searched against the SwissProt_2022 database (Homo sapiens, 20402 entries) using Mascot (Matrix Science, London, UK; version 2.8.0.1). The search variables that were used were: 10 ppm mass tolerance for precursor ion masses and 0.5 Da for product ion masses; digestion with trypsin; a maximum of two missed tryptic cleavages; variable modifications of oxidation of methionine and phosphorylation of serine, threonine, and tyrosine; 13C=1. The resulting Mascot .xml file was then imported into Progenesis, allowing for peptide/protein assignment, while peptides with a Mascot Ion Score of <25 were not considered for further analysis. Protein quantification was performed using only non-conflicting peptides and precursor ion-abundance values were normalized in a run to those in a reference run (not necessarily the same as the alignment reference run). Unbiased hierarchal clustering analysis (heat map) was performed in Perseus.^76–77

Abbreviations

BET

bromodomain and extra terminal

BRD4

bromodomain and extra terminal

DMEM

dulbecco’s modified Eagle’s medium

HPLC

high-performance liquid chromatography

POI

proteins of interest

PROTAC

proteolysis targeting chimera

SA-β-gal

senescence-associated β-galactosidase

SASP

senescence-associated secretory phenotype

TGI

tumor growth inhibition

UPLC

ultra performance liquid chromatography

VHL

von hippel-lindau