The BASHY Platform Enables the Assembly of a Fluorescent Bortezomib–GV1001 Conjugate

Abstract

In this study, we show that fluorescent boronic-acid derived salicylidenehydrazone complexes (BASHY) can function as fluorescent linkers for bioconjugates that were used to monitor the delivery of the proteasome inhibitor bortezomib (Btz) to HT-29 cancer cells. BASHY complexes were structurally optimized to improve the stability of the complex in buffered conditions (ammonium acetate, pH 7 up to t_1/2 = 40 h), photophysically characterized regarding their fluorescence properties and used in confocal microscopy colocalization studies that revealed their intracellular sequestration by lipid droplets. The accumulation in these hydrophobic organelles limited the hydrolysis of the complex and consequently the drug release, a problem that was circumvented by the conjugation of the BASHY-Btz complex with a cell-penetrating peptide GV1001-C. The conjugate exhibited an improved cytoplasmic availability as confirmed by confocal fluorescence microscopy studies and an improved potency against HT-29 cancer cells (IC₅₀ = 100 nM) as compared to the nontargeted complex (IC₅₀ = 450 nM).

Benefiting from important developments in human cancer biology, chemotherapy evolved in recent years to a curative state for different subsets of cancers. This success is intimately related with the optimization of existing drugs and drug regimens, but also with the discovery of innovative therapeutics such as bioconjugates. These multifunctional constructs, combining the efficiency of potent cytotoxic drugs with the targeting ability of specific biomolecules, emerged as lead therapeutics in oncology, with several representatives reaching the market.¹ Despite conceptually being relatively simple, the construction of these multifunctional constructs is often hampered by the complexity of the linker technology. As a basic requirement, the linker is required not only to connect both functional components (i.e., the drug itself and the biocompatible vehicle), but also to ensure the bioconjugate’s stability in circulation and the effective drug release at the site of action.² In addition to the obvious synthetic hurdles, the development of such platforms also requires the precise characterization of intracellular trafficking of the bioconjugate, using fluorescence microscopy. To enable such studies, bioconjugates need to be functionalized with fluorescent probes, which is typically accomplished either by replacing the cytotoxic drug or by integrating fluorophores in the linker structure.³ Both approaches are not ideal, because in the first case, the bioconjugate does not resemble the functionality of the cytotoxic construct, and in the latter, the structural modification of the linker may compromise the bioconjugate’s stability and delivery mechanisms. Having these shortcomings in mind, an alternative design to expedite the construction and intracellular monitoring of bioconjugates would be the engineering of linkers that are inherently fluorescent. Beside an ideally straightforward synthetic access, these could be designed to offer the necessary functional assets for effectively controlling the stimuli-responsive and thereby selective on-site delivery of the therapeutic drug.

Recently, we showed that boronic acids (BAs) enable the modular assembly of cleavable drug conjugates.⁴ In addition to this, we also demonstrated that BAs can be used as a configurational lock for the construction of fluorescent boronic acid derived salicylidenehydrazone complexes (BASHY), which exhibit suitable properties to be used in bioimaging applications.⁵ Taking these studies in consideration, we envisioned that a straightforward route to develop fluorescent cytotoxic bioconjugates would be the reengineering of the BASHY platform to work as a fluorescent targeting drug conjugate (Figure 1).

Figure 1.

Reengineering the BASHY platform to work as a fluorescent linker for targeting drug conjugate.

To test this idea, we selected BASHY 4 as the lead architecture. This dye was prepared (Figure 2 and Supporting Information (SI), section 1.5) and used in the labeling of HT-29 cancer cells. This process was monitored by confocal fluorescence microscopy, which revealed that 4 (λ_abs,max = 425 nm, λ_fluo,max = 504 nm, Φ_fluo = 0.10 in toluene) accumulates at the intracellular level in the same organelles that are also labeled by lipid droplet (LD) marker Nile Red (Pearsons’s correlation coefficient, PCC = 0.850), while exhibiting a very poor colocalization (PCC = 0.094) with LysoTracker staining (Figure 2 and SI, Figure S34).⁶ This strongly suggests that 4 is accumulating in lipid droplets (LDs) in HT-29 cancer cells. However, kinetic studies performed on this system showed that overtime the fluorescence decays, suggesting that the dye is partioning between the LDs lipid phase and the cytoplasm where it possibly undergoes hydrolysis (see SI, section 4.8).

Figure 2.

Synthesis of BASHY 4, optical spectra (UV/vis absorption in black and fluorescence in red) and confocal fluorescence microscopy colocalization assays of 4 with lysosome marker LysoTracker Deep Red (red) and Nile Red (red) for lipid droplet staining (red) in HT-29 cells. In both assays, the nucleus was labeled with marker Hoechst 33342 (blue).

In the next step, the same salicylidenehydrazone ligand that was used in 4 was applied in the complexation of bortezomib (Btz), which is a BA-derived cytotoxic drug, widely applied as a proteasome inhibitor.⁷ As shown in Figure 3 and in SI, section 1.6, the one-pot reaction readily afforded BASHY 5. This compound was shown to be fluorescent (λ_abs,max = 404 nm, λ_fluo,max = 481 nm, Φ_fluo = 0.08 in toluene), and it displayed significant stability in ammonium acetate at pH 7 (t_1/2 = 13 h; SI, Chart S1) and a higher partition coefficient (SI, Chart S3) for the lipid phase (K_p POPC/water = 38310 ± 6754) as compared to BASHY 4 (K_p POPC/water = 9575 ± 1406).

Figure 3.

Synthesis of BASHY 5, optical spectra (UV/vis absorption in black and fluorescence in red), cytotoxicity, and confocal fluorescence microscopy localization assays of 5 in HT-29 cells. The BASHY 5 in green and the nucleus labeled with marker Hoechst 33342 (blue). The cell viability of HeLa cells incubated with Btz (solid line) or BASHY 5 (dashed line) was determined with MTT assay (see SI, section 4.2).

Once internalized, BASHY 5 was expected to hydrolyze and release its cytotoxic payload. Therefore, the cytotoxicity of the complex and Btz were evaluated against HT-29 cancer cells. In these assays, Btz exhibits an IC₅₀ of 3.2 nM, while 5 is characterized by an IC₅₀ of 19.5 nM. Similar results were seen in HeLa cells (Btz – IC₅₀ = 20 nM; 5 – IC₅₀ = 134 nM). However, under dilute aqueous conditions, we observed that 5 hydrolyzed (see SI, section 2.2) and released the Btz drug. This instability certainly contributed to the observed cytotoxicity, although, by confocal microscopy we observed that like 4, complex 5 internalizes and accumulates in LDs (Figure 3 and SI, Figure S35). Therefore, the decrease in potency in both cell lines was rationalized with the increased hydrophobicity of 5 as compared to Btz, which favored the intracellular sequestration of the complex by LDs (Figure 3).

To study if the LD organelles had an impact on 5 and Btz potency, HT-29 and HeLa cancer cells were supplemented with oleic acid to induce intracellular LD accumulation.⁸ HeLa cells treated with oleic acid exhibit a pronounced increase in the total volume of lipid droplets, which was confirmed by confocal fluorescence microscopy (Figure 4 and SI, Figure S39). Therefore, this cell line was selected as a model to perform this evaluation. In these conditions the potency of Btz against HeLa cells (IC₅₀ = 20 nM without oleic acid supplementation) was considerably reduced (IC₅₀ > 200 nM), and similarly, the potency of 5 in these supplemented conditions was also decreased to an IC₅₀ of >200 nM. The general character of this observation was validated with doxorubicin and SN38, which, despite exhibiting distinct mechanisms of action, also displayed a much lower potency in the HeLa cells with increased concentration of LDs (Doxo – IC₅₀ ∼ 90 nM vs >1 μM; SN38 – IC₅₀ 349 nM vs >1 μM). These results, indicate that these hydrophobic organelles can indeed sequester hydrophobic molecules, inhibiting their cytotoxic effect.⁹

Figure 4.

LD cell content upon supplementation HeLa and HT-29 cells with oleic acid. LDs were imaged by confocal fluorescence microscopy using Nile Red. Quantification of LD number and occupation percentage of total available area per cell was carried out as described in SI, section 4.4.

Considering these results, the BASHY core was modified by substitution of the imine with a phenyl moiety. This was expected to improve hydrolytic stability of the complex by screening the electrophilic center from the reaction with water (Figure 5 and SI, section 1.7). Considering this design, 7 was readily synthesized from 6 in 68% yield. Gratifyingly, this modification produced a more stable complex 7 (t_1/2 = 40 h, SI, Chart S2), with fluorescence characteristics that were favorable to perform bioimaging (λ_abs,max = 412 nm, λ_fluo,max = 490 nm, Φ_fluo = 0.02 in toluene). Regarding the cytotoxicity against HT-29 cancer cell line, the modification favored the stability of the complex but also contributed to a considerable decrease in potency (IC₅₀ = 450 nM), while maintaining the partition into LDs (Figure 5 and SI, Figure S36).

Figure 5.

Synthesis of BASHY 7, optical spectra (UV/vis absorption in black and fluorescence in red), cytotoxicity, and confocal fluorescence microscopy localization assays of 7 in HT-29 cells. The BASHY 7 in green and the nucleus labeled with marker Hoechst 33342 (blue). Cell viability upon treatment with BASHY 7 was determined using the MTT assay (see SI, section 4.2).

Next, the BASHY core 7 was equipped with a PEG-N₃ motif to improve the aqueous solubility of the complex and to enable the core postfunctionalization. As shown in Figure 6 and SI, section 1.8, BASHY 8 was obtained in 56% isolated yield, and the azide function was then used to install a cyclooctyne–PEG–maleimide linker 9 through a strain-promoted azide–alkyne cycloaddition (SPAAC). This reaction generated the intermediate 10 (Figure 6 and SI, section 1.9), which was subsequently employed in the bioconjugation with GV1001 peptide, engineered at the C-terminal with a cysteine residue (GV1001-C). The GV1001 peptide was selected for this study as it can act as a cell-penetrating peptide that facilitates the transport of molecular cargo across the plasma membrane.¹⁰ The formation of bioconjugate 11 (Figure 6 and SI, section 1.10) was confirmed by high-resolution mass spectrometry (SI, Figures S31and S32). The enhanced stability of this core enabled the conjugate purification by semipreparative reversed-phase HPLC.

Figure 6.

Synthesis of BASHY 10 by SPAAC with 9. Deconvoluted mass spectrum of the bioconjugation reaction with GV1001-C peptide in ammonium acetate 20 mM pH 7/DMF (31% yield). Cytoxicity and confocal fluorescence microscopy localization assays of 11 in HT-29 cells. The BASHY 11 in green and the nucleus labeled with marker Hoechst 33342 (blue). Nile Red (red) was used as a LD marker and LysoTracker Deep Red as a lysosome marker. Cell viability upon treatment with BASHY 11 was determined using the MTT assay (see SI, section 4.2). Compounds 10 and 11 are a mixture of isomers resulting from the SPAAC reaction, only one isomer is presented for clarity.

Once prepared, the intracellular trafficking of the bioconjugate was studied by confocal fluorescence microscopy using HT-29 cancer cells. The bioimaging experiments evidenced a poor colocalization (Figure 6 and SI, Figure S37) with either LysoTracker Deep Red (Pearson’s correlation coefficient PCC = 0.197) or Nile Red (PCC = 0.273). These observations correlate well with lower partition coefficient for the lipid phase (K_p POPC/water = 650 ± 718) and confirms that the functionalization of 7 with GV1001-C peptide inhibits the sequestration of 11 by LDs (see SI, Chart S3). This cytoplasmatic availability of the bioconjugate had also an impact on the complex potency and bioconjugate 11 exhibits an IC₅₀ of 100 nM, which is 4.5-fold higher than observed for BASHY 7. These results validated the fluorescent BASHY complexes as suitable linkers to design targeting drug conjugates by using boronated cytotoxic drugs.

In summary, in this study, we modified fluorescent boronic acid derived salicylidenehydrazone complexes (BASHY) to be used as fluorescent linkers to deliver the proteasome inhibitor bortezomib to cancer cells, using HT-29 cells as model in a proof-of-concept approach. We observed that these hydrophobic fluorescent complexes were sequestered by lipid droplets, and this intracellular accumulation inhibited partially the delivery of the drug. The functionalization of the platform with the cell-penetrating peptide GV1001-C generated the bioconjugate 11. This conjugate has improved cytoplasmatic availability, as confirmed in confocal fluorescence microscopy studies and an improved potency against HT-29 cancer cells (IC₅₀ = 100 nM). This contrasts with the lower potency of the nontargeted complex 7 (IC₅₀ 450 nM), which was shown to strongly accumulate intracellularly in LDs.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.1c00615.

• Synthetic methods used for the preparation of the complexes as well as their structural characterization, stability, and photophysical evaluation and in vitro cellular assays (PDF)

Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

The authors declare no competing financial interest.

Special Issue

Published as part of the ACS Medicinal Chemistry Letters virtual special issue Medicinal Chemistry in Portugal and Spain: A Strong Iberian Alliance.