Antitumor Activity of 1,18-Octadecanedioic Acid-Paclitaxel Complexed with Human Serum Albumin.

We describe the design, synthesis, and antitumor activity of an 18 carbon α,ω-dicarboxylic acid monoconjugated via an ester linkage to paclitaxel (PTX). This 1,18-octadecanedioic acid-PTX (ODDA-PTX) prodrug readily forms a noncovalent complex with human serum albumin (HSA). Preservation of the terminal carboxylic acid moiety on ODDA-PTX enables binding to HSA in the same manner as native long-chain fatty acids (LCFAs), within hydrophobic pockets, maintaining favorable electrostatic contacts between the ω-carboxylate of ODDA-PTX and positively charged amino acid residues of the protein. This carrier strategy for small molecule drugs is based on naturally evolved interactions between LCFAs and HSA, demonstrated here for PTX. ODDA-PTX shows differentiated pharmacokinetics, higher maximum tolerated doses and increased efficacy in vivo in multiple subcutaneous murine xenograft models of human cancer, as compared to two FDA-approved clinical formulations, Cremophor EL-formulated paclitaxel (crPTX) and Abraxane (nanoparticle albumin-bound (nab)-paclitaxel).

Harnessing human serum albumin (HSA) as a drug carrier in oncology is an established strategy aimed to improve dosing, safety, and efficacy.[1−4] HSA is the most abundant serum protein with myriad functions, including a central role in metabolic pathways as a transporter of long-chain fatty acids (LCFAs) and hydrophobic molecules, both natural and synthetic.[5−7] Recently, HSA and its endogenous ligands have gained interest for their roles in dysregulated metabolism tied to tumor growth.[8,9] HSA is scavenged by solid tumors as a catabolic source of glutamine.[10,11] Further, exogenous LCFA uptake and metabolism are prominent in the metastatic progression of human cancers with poor prognoses,[12−15] as sustained tumor growth requires upregulation of nutrient acquisition. Therefore, drug-modified LCFAs that mimic the natural binding modes with HSA is a drug delivery strategy for cancer with potential in optimized dosing, stability, solubility, and efficacy by hitchhiking on these metabolic pathways.

Binding interactions between HSA and LCFAs have been extensively studied, with as many as 7 moderate- to high-affinity binding pockets identified through nuclear magnetic resonance (NMR) and protein crystallography.[16−19] In five of these sites, LCFA binding is stabilized through specific electrostatic interactions between the terminal LCFA carboxylate anion and positively charged amino acid residues at the base of otherwise hydrophobic binding pockets of HSA. These interactions are a consistent structural feature of thermodynamically stable, natural binding modes between LCFAs and HSA. However, in classical drug lipidation strategies, this electrostatic interaction is extinguished when the single carboxylic acid of an LCFA is conjugated to the warhead, often via ester, amide, or hydrazone linkage.[20−25] Indeed, the first and only example utilizing a long-chain fatty diacid, Ozempic (semaglutide), was approved by the FDA in 2017. Maintaining the naturally evolved LCFA carboxylate contact is a critical feature contributing to semaglutide’s performance as a once-weekly glucagon-like-peptide-1 receptor agonist for type-2 diabetes.[26]

Herein, we employ an 18 carbon α,ω-dicarboxylic acid as the HSA-binding motif of a prodrug conjugate. This conserves the high affinity electrostatic interactions between the ω-carboxylate and HSA, while maintaining the α-carboxylate for drug conjugation. This approach preserves the natural binding mode and mimics HSA engagement of natural LCFAs (Figure ). Specifically, 1,18-octadecanedioic acid (ODDA) was conjugated to paclitaxel (PTX) to produce a monofunctionalized PTX ester at the 2′-hydroxyl (ODDA-PTX, Figure a, Figures S1 and S2). PTX is one of the most widely used chemotherapeutic agents. PTX promotes tubulin polymerization and arrests cell division, leading to apoptotic cell death.[27] Functionalization of PTX at the 2′-hydroxyl prevents this mode of action.[28] Indeed, an in vitro polymerization assay confirmed ODDA-PTX performed as a prodrug (Figure S3). Therefore, cytotoxic activity of ODDA-PTX is regained only when PTX is free from ODDA.

Figure 1

Design of ODDA-PTX and HSA binding. (a) Chemical structure of ODDA-PTX. (b) Modeling of ODDA-PTX binding to 7 LCFA sites on HSA. (c) Molecular model of HSA with ODDA-PTX bound in native LCFA site 4 (sphere). (d) NMR titration of ODDA-PTX into a pre-equilibrated, saturated solution of palmitic-1-13C acid (C16:0, 2.3 mM) with HSA (0.45 mM). ODDA-PTX displaces C16:0 in all observable fatty acid binding pockets at concentrations lower than that of the native ligand, demonstrating a high binding affinity to HSA.

ODDA-PTX was mixed with HSA to yield a water-soluble complex at a 5:1 prodrug:protein mol ratio, designated VTX (Figure b,c. “V” = Roman numeral for five, signifying the number of equivalents of ODDA-PTX per HSA. “TX” = taxane). VTX was readily reconstituted in PBS following lyophilization, yielding a homogeneous solution. HSA binding sites for palmitic acid (C16:0) and other LCFAs are known;[16,17] thus the interaction of ODDA-PTX with HSA was examined experimentally by NMR displacement studies[18,19] with palmitic-1-13C acid (Figure d). ODDA-PTX was titrated into a pre-equilibrated solution of HSA with palmitic-1-13C acid, showing its rapid displacement.

To corroborate these data, VTX was modeled using molecular dynamics simulations (Figure b,c, Figures S4–S6, Table S1). It was determined that fatty acid binding sites 1, 2, 4, and 5 were the most likely binding locations for ODDA-PTX; modes consistent with native binding of LCFAs. Importantly, VTX preserves the tertiary structure and thermal stability of HSA as determined by circular dichroism spectroscopy (Figure S7). A rapid equilibrium dialysis assay showed that >95% of ODDA-PTX was bound to HSA down to 0.5 μM (Table S2), with isothermal titration calorimetry confirming multiple binding sites (Figure S8). Dynamic light scattering and cryogenic transmission electron microscopy revealed VTX as a fully dispersed prodrug:protein complex, similar to native HSA, not a nanoparticulate aggregate (Figure S9).

The cell-free stability of ODDA-PTX and VTX was assayed over a range of conditions by high performance liquid chromatography (Figures S10–S12). Cytotoxicity of VTX was assessed in vitro, confirming activity at nM concentrations with respect to PTX (Table S3). The metabolic stability and in vitro intrinsic clearance of ODDA-PTX was compared to PTX in human liver microsomes (Figure S13), demonstrating similar rates of metabolism. Significantly, blocking the fatty acid translocase/CD36 receptor, a key protein involved in uptake of LCFAs across cell membranes[29] with implications in metastatic cancers,[15,30] reduced the cytotoxicity of ODDA-PTX, but not PTX (Figure S14). This provides evidence that ODDA-PTX enters cells through different pathways from PTX, and importantly, through the same pathways as LCFAs.

VTX was evaluated as an antitumor agent in subcutaneous xenograft mouse models of human cancer. Therapeutic efficacy was determined by monitoring tumor growth and survival versus nontreated controls and compared to two FDA-approved paclitaxel formulations, Abraxane (nanoparticle albumin-bound PTX) and Cremophor-formulated PTX (crPTX). Initial studies were carried out in the HT-1080 human fibrosarcoma model, as it is predictable, fast-growing, and known to be taxane-responsive (Figure ).[31,32] The therapeutic index (TI) was defined as the ratio of maximum tolerated dose (MTD) to minimum effective dose (MED),[33] where MTD was the highest nonlethal deliverable dose and less than 20% weight loss,[34] and MED was the lowest dose showing a delay in tumor growth relative to nontreated controls.[35] Efficacy of VTX was investigated in the HT-1080 model following intravenous (IV) injection at 5, 60, 120, and 250 mg/kg with respect to PTX concentration every 7 days for 4 weeks (q7dx4). Animals responded to VTX in a dose-dependent manner (Figure a). Complete tumor regression (5/8) or dramatically suppressed progression was observed in animals administered VTX at 250 mg/kg, with no evidence of drug-associated toxicity observed (Figure b, Figures S15 and S16). Conversely, animals administered Abraxane at 20 mg/kg with respect to PTX concentration experienced significant treatment-associated lethality (Figure S17). Therefore, the highest dose of Abraxane administered in this model was limited to 15 mg/kg. We note that these MTDs are lower than what is reported in several published studies.[36−38] Similar antitumor activity of crPTX at 15 mg/kg was observed to that of Abraxane at 15 mg/kg (Figure c,d), the reported MTD of crPTX.[39]

Figure 2

Efficacy and survival in HT-1080 fibrosarcoma model (n = 8 per group). Doses were administered IV via tail vein (q7dx4). (a) Dose–response of VTX with respect to PTX concentration, as compared to nontreated animals. Tumor volume is relative to initial volume on first day of treatment. Efficacy graphs for each group are truncated once survival drops below 50%. (b) Corresponding survival plot to panel a. (c) Relative tumor growth for VTX versus Abraxane and crPTX at MTD. (d) Corresponding survival plot to panel c. In this model, doses of Abraxane could not exceed 15 mg/kg due to toxicity (see SI).

Elevated doses of VTX significantly extended survival time and inhibited tumor growth over nontreated controls (Figure b and Figure S18), whereas the improvement of survival time for Abraxane and crPTX was modest. The lowest administered dose (5 mg/kg) of VTX and Abraxane had similar effects on median survival time and both treatments sufficiently suppressed tumor growth relative to nontreated controls over the course of the study (p-value = 0.041 and 0.045, respectively, based on an unpaired t test where p < 0.05 was considered statistically significant). This was defined as the MED for both VTX and Abraxane in this model and with this dosing regimen (Figure S19). In evaluating the efficacy in tumor growth suppression and survival of animals treated with VTX and Abraxane, we conclude a distinct advantage of VTX over the FDA-approved drug. In this model, VTX had a MED of 5 mg/kg and a MTD of >250 mg/kg with respect to PTX concentration. Higher doses could not be explored due to solubility constraints of HSA used in the formulation of VTX. Thus, the TI of VTX is >50 in the HT-1080 model at a q7dx4 dosing regimen. Conversely, Abraxane-treated animals experienced significant toxicity at all doses exceeding 15 mg/kg (see Figure S17), revealing a TI of 3 for Abraxane.

Pharmacokinetic (PK) parameters of VTX and Abraxane were determined in both healthy and tumor-burdened (HT-1080 xenograft) nu/nu mice and calculated using standard noncompartmental methods (Table ).[40] At 1 h postinjection, both the metabolite PTX and ODDA-PTX were found in plasma samples from animals administered VTX. However, ODDA-PTX is present at a 5-fold higher concentration than PTX. Maximum plasma concentration (Cmax) of PTX in animals administered Abraxane at 20 mg/kg is similar to those administered VTX at 250 mg/kg (Figures S20 and S21). Half-life and area under the plasma concentration versus time curve (AUC) of PTX in animals treated with VTX is higher than that of Abraxane, suggesting that VTX acts as a drug depot for the cytotoxin.

Table 1

Calculated PK Parameteres of Paclitaxel Released from VTX and Abraxane Following Single IV Injection in HT-1080 Murine Xenograftsa,b

Group	Cmax (ng/mL)	AUC ((h·ng)/mL)	Vdist (mL)	t1/2 (h)
VTX	508 ± 188	4688 ± 797	1257 ± 374	9.80 ± 1.20
Abraxane	1607 ± 469	5640 ± 469	481 ± 163	4.39 ± 0.62

Animals dosed at 20 mg/kg with respect to paclitaxel content of each formulation.

Detected entity is intact paclitaxel.

Biodistribution of VTX dosed at 250 mg/kg with respect to PTX content in HT-1080 xenografts was evaluated in comparison to Abraxane and crPTX in both the tumor and the liver at 1, 4, and 24 h postinjection (Figure ). The highest initial tumor concentration was in animals administered crPTX (Figure a). However, the tumor concentration of metabolite PTX in animals administered VTX steadily increased as a function of time, consistent with the depot-like PK data. Minimal accumulation in the liver was observed for VTX, compared to the high levels observed for both the nanoscale aggregates of Abraxane (see Figure S9) and crPTX at early time points (Figure b). It has been proposed that Cremophor imparts nanoparticle-like properties to the crPTX formulation[41] with liver accumulation as a result.

Figure 3

Biodistribution of PTX (n = 3 per group). Concentration of PTX in HT-1080 xenografts at 1, 4, and 24 h postinjection with either VTX, Abraxane, or crPTX in (a) tumor and (b) liver.

Given the promising efficacy and tolerance (as measured by body weight) observed at 250 mg/kg with respect to PTX concentration, the toxicity of VTX at this dose in nu/nu mice was further evaluated following IV injection (q7dx4) and compared to crPTX and Abraxane at 15 mg/kg. A complete blood count (CBC) panel and serum chemical analysis was obtained 24 h after the final injection (Table S4, Figures S22 and S23). Nearly all parameters were equivalent between all treatment groups, despite animals being administered VTX at a 16-fold higher dose of PTX than those administered Abraxane or crPTX (Figures S24 and S25).

Additional efficacy studies were completed in HT-29 (human colorectal adenocarcinoma)[3] and PANC-1 (human pancreas ductal carcinoma) models, both known targets of taxane therapy (Figure ).[3,42] The trends in efficacy from the HT-1080 model were also observed in these models, where VTX at 250 mg/kg was well-tolerated and flat-lined tumor growth (Figure , Figure S26). Abraxane showed no significant therapeutic effect at 15 mg/kg in the PANC-1 model. Thus, we explored Abraxane at a higher dose (30 mg/kg) with 2/8 mice experiencing acute lethality (Figure S27). Conversely, 2 animals experienced complete tumor regression upon administration of high doses of VTX.

Figure 4

Efficacy of VTX (n = 8 per group) in xenograft models. (a) Pancreatic carcinoma PANC-1, and (b) colorectal adenocarcinoma HT-29.

In conclusion, we have developed a carrier system for albumin-mediated delivery of PTX. The design strategy is based on mimicking the natural interaction between HSA and LCFAs, by monofunctionalizing ODDA with PTX. We have demonstrated efficacy of VTX in three xenograft models: HT-1080 (fibrosarcoma), PANC-1 (pancreatic cancer), and HT-29 (colon cancer). The overall survival of animals is significantly extended when administered VTX, compared to clinical formulations. The PK profile of VTX differs significantly from Abraxane, as does the biodistribution compared to Abraxane and crPTX. Given the relationship between HSA, fatty acids, and the tumor microenvironment, this is a potential strategy for drug delivery to cancers that have increased demands for HSA and fatty acids.[43,44] This may be of importance in cancers that demonstrate multidrug resistance, where conventional PTX treatment is ineffective.[45,46] We propose this strategy as a generalizable new route for the facile modification of drugs for engaging HSA without the need for complex nanoscale formulations or covalent modification of the protein.[47]