|Home | About | Journals | Submit | Contact Us | Français|
Developing selective strategies to treat metastatic cancers remains a significant challenge. Herein, we report the first antibody-recruiting small molecule (ARM) that is capable of recognizing the urokinase-type plasminogen activator receptor (uPAR), a uniquely overexpressed cancer cell-surface marker, and facilitating the immune-mediated destruction of cancer cells. A co-crystal structure of the ARM-U2/uPAR complex was obtained, representing the first crystal structure of uPAR complexed with a non-peptide ligand. Finally, we demonstrated that ARM-U2 substantially suppresses tumor growth in vivo with no evidence of weight loss, unlike the standard-of-care agent doxorubicin. This work underscores the promise of antibody-recruiting molecules as immunotherapeutics for treating cancer.
Tumor metastasis involves the invasion of cancer cells into surrounding tissues, a process often accelerated by cell- surface proteases. One such protease, the urokinase-type plasminogen activator (uPA), breaks down extracellular matrix proteins and activates migration-inducing signal cascades upon binding to the urokinase-type plasminogen activator receptor (uPAR). Both uPA and uPAR expression are substantially upregulated in invasive cancer cells com- pared to healthy cells or benign tumors. In clinical settings, uPAR levels have been shown to correlate with metastatic potential and poor clinical outcomes.[1a] As such, uPAR has gained recognition as a promising target for treating meta- static cancers from diverse tissues of origin, including breast, colon, stomach, and bladder.[3,4]
Antibody-recruiting molecules (ARMs) are bifunctional molecules capable of delivering endogenous antibodies to disease-causing entities, leading to their destruction and/or clearance by the immune system. Our group and others have previously developed ARMs that target various cancers and infectious agents. These novel immunotherapies have the potential both to complement protein-based agents while overcoming their challenges, such as poor oral bioavailability, high molecular weights, and immunogenicity.
Our group previously reported an ARM capable of targeting uPAR-expressing cancer cells for immune-mediated cell death. Although effective in vitro, the reported con-struct (termed ARM-U1) contained the uPA protein at its target binding terminus (TBT), imparting many of the limitations of biologics. We therefore hypothesized that the therapeutic potential of this agent could be significantly improved upon by replacing the uPA protein with a high- affinity uPAR-binding small molecule (Figure 1 A).
Herein, we report the design, synthesis, and in vitro and in vivo evaluation of a second-generation, low-molecular- weight (< 1000 amu) ARM derivative termed ARM-U2. In place of uPA, ARM-U2 incorporates a restructured analogue of IPR-803, a uPAR inhibitor identified previously using a virtual screen, into its TBT. ARM-U2 targets the uPA binding site on uPAR with low nanomolar affinity, induces immune-mediated phagocytosis and cytotoxicity of uPAR- expressing cells in culture, and also inhibits tumor progression in vivo, possessing comparable efficacy to the standard-of- care agent doxorubicin but without the substantial weight loss associated with doxorubicin treatment. We also report a co- crystal structure of the ARM-U2/uPAR complex, which to our knowledge represents the first uPAR/small-molecule inhibitor complex to appear in the literature. This work underscores the promise of small-molecule immunotherapeutics for treating patients with uPAR-expressing metastatic cancers.
In designing ARM-U2, we first sought to identify a site in the IPR-803 scaffold at which to affix a linker connecting the TBT to a 2,4-dinitrophenyl (DNP) antibody-binding terminus (ABT). To this end, we performed computationally assisted docking experiments, which suggested that functionalization of the IPR-803 core at position 2 of ring C with triethylene glycol derived PEG-3 linker 9 would be sufficient to drive solvent exposure of the ABT (1; Figure 1 B). Also, computational studies performed previously and in our laboratory suggested that electrostatic interactions between R53 and the carboxylate of IPR-803 are critical for driving ligand affinity. We therefore hypothesized that repositioning this carboxylate function closer to R53,and/or incorporating additional acidic sites into the ligand, would improve its overall binding affinity. To test this, we performed further docking studies with model derivatives differing in the number or position of negatively charged substituents within ring D. We explored biscarboxylate 15, extended monocarboxylate 16, and the mono- and bissulfonate deriv- atives 17 and 2, respectively (Figure 2). Compound 15 was predicted to bind uPAR with comparable affinity to 1, whereas the other derivatives were calculated to have significantly stronger docking scores, likely both because of stronger electrostatic contacts and additional bonding interactions with the hotspot residues R53 and K50 (Table 1, Figure S1).
Informed both by these computational results and published SAR studies, we prepared several ARM-U2 derivatives for further evaluation (Figure 2). These were synthesized in a divergent manner starting from isoxazole 3 (Figure 2, Scheme S2).[8b] Thus, exposure of 3 to a diverse set of commercially available functionalized anilines under Lewis acid catalysis at ambient temperature led to the chemo- selective displacement of the 4-bromo substituent to afford variously substituted intermediates 4–8 in high yields. Treat- ment of 4–8 with amine base and (S)-3-piperidinecarboxylate derivatives 9–12, containing linkers of various lengths and antibody-binding motifs, led to a second displacement process of the bromine atom in the 2 position. Ester deprotection where appropriate yielded products 1, 2, and 13–20 in good to excellent yields.[8b]
We then evaluated the binding of compounds 1, 2,and 13–20 to uPAR using a competition ELISA protocol (Table 1). We first observed that the KI and Kd values determined for 1 were in close agreement with the reported 200 nm affinity of IPR-803 for uPAR (see also Figure S2).[8a] This result supports computational predictions that linker ABT installation at position 2 of ring C would minimally perturb uPAR binding affinity. ARM-U candidates (1, 2, and 13–20) were also found to bind uPAR competitively with uPA possessing affinities ranging from low double- to triple-digit nanomolar values (Table 1, Table S1, and Figure S2).Although the absolute magnitudes of the computational docking energies did not correlate with experimental binding affinities, the highest ranked compound, bissulfonate 2, was found to possess the highest uPAR binding affinity with a KI value of 12 nm (Table 1). These binding results validated our hypothesis that repositioning and/or increasing the negative charge on the D ring of ARM-U2 derivatives could afford significant increases in uPAR binding affinity.
We next evaluated the ability of 1 and 2 (henceforth termed “ARM-U2”) to induce responses from immune effector cells. We measured the antibody-dependent cellular phagocytosis (ADCP) of uPAR-expressing A172 glioblas- toma cells by IFNg-activated U937 monocytes using a two- color flow-cytometry procedure. Both 1 and ARM-U2 exhibited concentration-dependent phagocytosis with max- imum efficacy observed at approximately 1 mm and 100 nm, respectively (Figure 3 A, Figure S4). The bell-shaped, auto- inhibitory dose-response curve observed at ARM-U2 con- centrations greater than 100 nm has been termed the “pro- zone” effect, and is expected for agents that function through ternary complexes, as has been described in detail previously. When present at high concentrations, unbound bifunctional agents compete with both antibody and cell- surface interactions in the ternary complex, driving the equilibrium to constituent binary complexes, thus preventing phagocytosis.
The dependence of ADCP on the presence of cell-surface uPAR was next assessed through competition experiments. Competition with soluble uPAR (suPAR)or 20, an ARM-U derivative lacking the DNP function, substantially suppressed ARM-U2-mediated phagocytosis (Figures 3 A, S4, and S6).
ARM-U2 was also capable of eliciting antibody-dependent cellular cytotoxicity (ADCC), which was suppressed in the presence of competitors (Figure S4). Taken together, these results support that ARM-U2 is capable of inducing both ADCP and ADCC in a manner dependent on binding anti-DNP antibodies and cell-surface uPAR.
To further assay the effector cell activation by ARM-U2, we investigated the release of pro-inflammatory cytokines by U937 monocytes in the context of ADCP assays. Thus, cellular supernatants were isolated from ADCP assay conditions and evaluated using a multi-inflammatory cytokine ELISA protocol.
Higher levels of IL-8 secretion were observed in assays containing 50 nm or 1 mm ARM-U2 compared to control samples lacking ARM-U2 (Figure 3 B). We believe that the slight decrease in IL-8 release at 1 mm versus 50 nm of ARM-U2 is another manifestation of the prozone effect, and coincides well with the concentration dependence observed in the ADCP experiments shown in Figure 3 A. These results are consistent with previous reports demonstrating that IL-8 is among the first cytokines released by monocytes and macrophages during the innate immune response, and serves as a downstream indicator of antibody-dependent cell-surface clustering and activation of CD64 (FcgRI). Also, the apparent absence of other cytokines in these experiments likely results from our exclusive use of CD64-expressing monocytic effector cells.
To gain insight into the ARM-U2 binding mode, we obtained a co-crystal structure of ARM-U2 bound to the uPAR soluble extracellular domain (suPAR). These studies revealed that ARM-U2 binds in a deep hydrophobic pocket within the uPA binding site on uPAR, and that the anthra- quinone core (rings A–C) is engaged in hydrophobic inter- actions with residues L55, L123, L150, L168, V125, and A255 (Figure S7, Table S2). The two D ring sulfonates in ARM-U2 are positioned within 2.8 and 3.5 Ł of R53 and T66, respectively, suggesting the presence of close-range electro- static and perhaps hydrogen-bonding contacts (Figure 4 A, B). These results suggest that enhanced polar contacts between the sulfonates present in ARM-U2 and hot-spot residues on uPAR are responsible for high-affinity binding.
Finally, we examined the efficacy of ARM-U2 in a B16- uPAR mouse allograft model developed in our laboratory. Mice were first immunized to produce anti-DNP IgG anti- bodies (Figure S8), then engrafted with B16-uPAR tumors as detailed in the Supporting Information. Animals received daily treatment with the indicated agents by intraperitoneal (IP) injection and were monitored for tumor growth. Mice treated with ARM-U2 (20 mpk or 100 mpk) showed significant decreases in the rate of tumor growth relative to those injected with PBS only, corresponding to a calculated tumor growth inhibition (TGI) of approximately 90 % (Figure 5 A). This level of TGI was comparable to mice treated with doxorubicin (Dox, 1 mpk), an antimitotic chemotherapeutic agent often used as a standard of care in treating metastatic malignancies, or a combination of ARM-U2 and Dox (20 mpk and 1 mpk, respectively). Furthermore, following treatment cessation, tumor regression persisted in all treatment groups for significantly longer than in the PBS control group (Figure 5 B). Specifically, median survival durations of between 27 and 37 days were observed for mice treated with Dox, ARM-U2, or a combination of these agents, whereas a median survival of 17.5 days was recorded for animals in the PBS control group. Remarkably, mice treated with ARM-U2 at either 20 or 100 mpk doses did not lose weight throughout the course of the experiments, in contrast to significant levels of weight loss observed among Dox- treated animals (Figure 5 C). Although lack of weight loss is not a definitive indicator of compound safety, these results represent a positive indication that ARM-U2 might possess an improved side-effect profile compared to doxorubicin.
In conclusion, we have developed a promising strategy for cancer immunotherapy that enables both the disruption of the oncogenic uPA–uPAR interaction and the ability to target uPAR-expressing cancers for immune-mediated cell death.
Although aryl sulfonates are uncommon in medicinal chemistry, we believe that the high negative charge of ARM-U2 plays an important role in its efficacy and potency, both in cellular assays and in vivo. Indeed, crystallographic studies indicate key interactions between sulfonic acid motifs in ARM-U2 and binding-site residues within uPAR lending both affinity and specificity to the interaction. Additional advantages of these charged groups might include reductions in off-target cellular binding and uptake, high aqueous solubility, and suppression of hydrophobic sticking and aggregation phenomena.
In vivo studies demonstrated that the tumor-killing efficacy of ARM-U2 was similar to that of doxorubicin, but without the weight loss associated with doxorubicin treatment. This observation may reflect an increase in selectivity associated with our targeted strategy as compared to traditional chemotherapeutics. Overall the ARM approach holds the exciting promise of combining the advantages of small molecule and biologic therapeutic modalities, along with the hope for providing better treatments for patients with cancer.
We would like to thank Dr. Julian Vastl, Dr. Tina Wang, and Dr. Dianna Bartel for helpful suggestions in the preparation of this manuscript. D.A.S. is a paid consultant for Bristol- Myers Squibb. This work was supported by funding from Bristol-Myers Squibb (OCR4997.11 to D.A.S.). We also acknowledge grants from the NRF (2013R1A2A2A01067638 and 2012M3C1A6035362 to A.E.C.).
Dr. Anthony F. Rullo, Department of Chemistry, Yale University 225 Prospect Street, New Haven, CT 06511 (USA)
Kelly J. Fitzgerald, Department of Pharmacology, Yale School of Medicine 333 Cedar Street, New Haven, CT 06520 (USA)
Dr. Viswanathan Muthusamy, Department of Chemistry, Yale University 225 Prospect Street, New Haven, CT 06511 (USA)
Min Liu, State Key Laboratory of Structural Chemistry Fujian Institute of Research on the Structure of Matter Chinese Academy of Sciences 155 Yang Qiao West Road, Fuzhou, Fujian 350002 (China)
Cai Yuan, State Key Laboratory of Structural Chemistry Fujian Institute of Research on the Structure of Matter Chinese Academy of Sciences 155 Yang Qiao West Road, Fuzhou, Fujian 350002 (China)
Prof., Dr. Mingdong Huang, State Key Laboratory of Structural Chemistry Fujian Institute of Research on the Structure of Matter Chinese Academy of Sciences 155 Yang Qiao West Road, Fuzhou, Fujian 350002 (China)
Minsup Kim, Department of Bioinformatics, Korea University 2511 Sejong-ro, Sejong, 339-700 (Korea)
Prof., Dr. Art E. Cho, Department of Bioinformatics, Korea University 2511 Sejong-ro, Sejong, 339-700 (Korea)
Prof., Dr. David A. Spiegel, Department of Chemistry, Yale University 225 Prospect Street, New Haven, CT 06511 (USA); Department of Pharmacology, Yale School of Medicine 333 Cedar Street, New Haven, CT 06520 (USA)