Pharmacokinetic and Biodistribution Studies of [¹⁸F]-Taccalonolide: A Covalent Microtubule Stabilizer with Antitumor Efficacy

Abstract

The taccalonolides are plant-derived microtubule stabilizers that covalently and specifically bind β-tubulin and provide antitumor efficacy in drug-resistant tumor models both in vitro and in vivo. Herein, we report the radiolabeling, in vitro uptake, and in vivo imaging of a ¹⁸F radiolabeled taccalonolide to investigate in vivo biodistribution, including accumulation in TNBC tumors. Biochemical and cellular studies demonstrate that fluorination does not alter biological activity and supports target engagement of the radiolabeled compound. In vivo studies in athymic nude mice bearing TNBC xenografts and healthy control animals show the highest uptake in the gallbladder and intestines with modest uptake in tumors. Unmodified free drug was detected in the gall bladder and intestines, suggesting hepatic transport of the compound prior to metabolism or irreversible target engagement. These results provide a strong rationale for the generation of novel derivatives and formulations to improve serum half-life and tumor targeting of the taccalonolides to ultimately provide new options for drug-resistant TNBC tumors.

Triple-negative breast cancer (TNBC) is a heterogeneous disease with limited expression of human epidermal growth factor receptor 2 (HER2), estrogen receptor (ER), and progesterone receptor (PR) and thus does not respond to therapies that target these receptors. TNBC represents 15–20% of all breast cancer diagnoses, with a prognosis significantly worse than that of other subtypes, particularly in the metastatic setting. Standard of care for TNBC often includes microtubule-targeted chemotherapeutics, including the taxane class of microtubule stabilizers. However, 30–50% of TNBC patients eventually evolve resistance to all currently approved microtubule targeted drugs, resulting in limited treatment options for this aggressive disease. Thus, there is a critical need for the development of microtubule targeted agents that retain efficacy in taxane-resistant settings.

The taccalonolides are a group of highly acetylated natural products isolated from plants of the genus Tacca that have demonstrated efficacy against taxane-sensitive and taxane-resistant cancer models in vitro and in vivo, including TNBC. − Similar to the taxane class of microtubule-stabilizing chemotherapeutics, the taccalonolides promote cytotoxicity by binding to and stabilizing microtubules, inhibiting their ability to segregate chromosomes during cell division, which results in mitotic arrest and apoptosis. , Semisynthetic approaches uncovered that epoxidation of the C22,23 alkene of the naturally derived taccalonolides is required for their specific covalent interaction with β-tubulin, resulting in cytotoxic activity against ovarian, breast, cervical, head and neck, and lung cancer models. − This irreversible target engagement by the taccalonolides allows them to retain efficacy in clinically relevant mechanisms of taxane resistance, including upregulation of efflux transporters, mutations in the taxane binding site, or increased expression of the βIII isotype of tubulin. − While some epoxytaccalonolides have demonstrated in vivo efficacy against TNBC tumors in vivo with systemic administration, , as a class they suffer from a narrow therapeutic index and rapid serum clearance.

Traditional drug binding and biodistribution studies of the epoxytaccalonolides have been complicated by their irreversible target engagement. A C6-fluorescently tagged epoxytaccalonolide provided a valuable tool to investigate target specificity in situ within cancer cells; however, this significant modification was not amenable for in vivo studies, as the fluorophore-tagged compound is susceptible to cleavage and not amenable for detection in vivo and would likely have significantly disrupted tissue distribution as compared to the parent compound due to the large modification. We proposed that radiolabeling the C7 hydroxyl of epoxytaccalonolide AJ with ¹⁸F would provide a unique pharmacological tool to further investigate the in vivo biodistribution and uptake in TNBC tumors of the epoxytaccalonolides. The ¹⁸F radionuclide is routinely used for molecular imaging of biologically active compounds due to its short half-life of 109.8 min and high positron branching ratio of 96.7% positron decay, which are ideal for tracing novel pharmaceuticals by positron emission tomography/computed tomography (PET/CT). , Fluorine also forms strong covalent bonds, making it suitable for use for the radiolabeling and in vivo imaging of small molecules, such as the epoxytaccalonolides. Herein we have developed a strategy for radiolabeling a mesylated taccalonolide analogue to produce 7-[¹⁸F]-fluorotaccalonolide AJ and investigated its biodistribution in healthy mice as well as accumulation in TNBC tumors.

Synthesis and Biological Characterization of Fluorinated Taccalonolides

Taccalonolide A was purchased from Chengdu Biopurify Phytochemicals, and its identity and purity were verified. The C15 acetate of taccalonolide A was hydrolyzed to generate taccalonolide B as previously described. Tacca B was then mesylated at the C7 position to give compound 1 (Scheme ). The mesylate was then displaced by potassium fluoride to afford the product 7α-deoxyfluorinated taccalonolide B (compound 2, Scheme ). Once fluorinated products were obtained and purified, their epoxidation at C22,23 with DMDO was monitored to completion to yield the desired epoxidized taccalonolide, compound 3 (Scheme , Supplemental Figure 1). We found that 3 retained microtubule stabilizing and antiproliferative activity similar to other epoxytaccalonloldes, including taccalonolide AJ, as evidenced by its ability to enhance tubulin polymerization in biochemical assays (Figure A), promote microtubule bundling in interphase cells (Figure B), and inhibit the proliferation of four different TNBC cell lines (Figure C). The minor 2–8-fold decrease in cellular potency with fluorination is consistent with the potency of other analogs, including taccalonolide AF that retains the same mechanism of action as a covalent microtubule stabilizer with in vivo antitumor efficacy. ,, Consistent with published data on the nonfluorinated C22,23 epoxytaccalonolides, we also found that 3 was irreversibly bound to microtubules formed in the biochemical polymerization assay (Supplemental Figure 3). In contrast, the fluorinated nonepoxidized compound, 2, did not covalently interact with tubulin, enhance its polymerization into microtubules, or promote any antiproliferative or cytotoxic effects in TNBC cell lines up to a concentration of 5 μM (Supplemental Figure 3), consistent with the well-described role of the C22,23 epoxide in the on-target biological activity of the taccalonolides. Taken together, these data suggest that fluorination at C7 does not change the mechanism of action of the epoxytaccalonolides as covalent microtubule stabilizers with efficacy against TNBC cells.

1. Synthesis of Precursor and Fluorinated Standard.

1.

Fluorinated taccalonolide retains microtubule stabilizing activity. (A) Polymerization of purified tubulin (20 μM) in the presence of vehicle or 10 μM paclitaxel, eribulin, taccalonolide AJ, or 3. Polymerization measured by absorbance at 340 nm upon warming to 37 °C. N = 3 ± SEM. (B) Immunofluorescence of microtubules in HCC1937 cells treated with vehicle (EtOH), 10 μM taccalonolide AJ, or 10 μM 3 for 4 h. Microtubules are shown in green, and DNA stained with DAPI is shown in blue. (C) Antiproliferative potency of taccalonolide AJ or 3 in four TNBC cell lines. Concentrations (nM) that inhibit proliferation by 50%. N = 3 ± SEM.

Radiolabeling

Upon confirming the bioactivity of cold fluorinated taccalonolide, we proceeded with ¹⁸F labeling using a similar approach. The precursor mesylated taccalonolide (1) was radiolabeled with an average approximate yield of 2% (range 1–4%) to generate nonepoxidized 7-[¹⁸F]­fluorotaccalonolide B ([ ¹⁸ F]-2) and then epoxidized to 7-[¹⁸F]-fluorotaccalonolide AJ [ ¹⁸ F]-3 (Scheme ) at sufficient levels to use as a radiotracer. Prior to utilizing [ ¹⁸ F]-3 in cell or animal studies, purity was shown to be >95% by both iTLC and HPLC (Supplemental Figure 2). Instant thin-layer chromatography shows the nonepoxidized and epoxidized radiolabeled products migrate fully with the ethanol (EtOH) solvent front, while unreacted ¹⁸F^– remains at the origin (Supplemental Figure 2). During HPLC analysis, unreacted ¹⁸F^– elutes early in the gradient, around 2.5 min, while the [ ¹⁸ F]-2 and [ ¹⁸ F]-3 elute at approximately 9.5 and 8.5 min, respectively. The nonradioactive fluorinated standard has approximately the same retention time as the nonepoxidized radiolabeled compound (9.5 min) with absorbance at 220 nm. The epoxidized standard 3 elutes with a retention time approximately one minute earlier than the unepoxidized standard 2 (Supplemental Figure 1).

2. Radiofluorination and Subsequent Epoxidation of Mesylate Precursor.

Cellular Uptake and Retention Studies

We next evaluated the cellular uptake and retention of both the epoxidized [ ¹⁸ F]-3, which should covalently and irreversibly bind tubulin, and the nonepoxidized precursor [ ¹⁸ F]-2 that lacks the epoxide moiety critical for covalent binding. Two TNBC cell lines (MDA-MB-231 and MDA-MB-468) were incubated with [ ¹⁸ F]-2 or [ ¹⁸ F]-3 for 30 min to 2 h, after which the percent of compound retained in cells was determined. Both cell lines showed a statistically significant increase in time-dependent uptake and/or retention of the epoxidized [ ¹⁸ F]-3 as compared to nonepoxidized [ ¹⁸ F]-2 (Figure and Supplemental Table 1). In a separate experiment, compounds [ ¹⁸ F]-2 and [ ¹⁸ F]-3 were added to cells for 1 h, after which the drug-containing medium was removed and replaced with fresh medium for an additional 15 min washout period to allow unbound drug to be effluxed from the cells. As expected, the C22,23-epoxidized compound [ ¹⁸ F]-3 was significantly retained after washout as compared to the nonepoxidized [ ¹⁸ F]-2. This difference between uptake and retention of the epoxidized and nonepoxidized radiolabeled compound is expected if functional irreversible target engagement via this moiety was retained in the fluorinated derivative and provides further confidence it can be used as a proxy for the unlabeled compound. These studies support that fluorination does not significantly alter the mechanism of action or irreversible C22,23 epoxide-mediated binding to β-tubulin.

2.

Intracellular uptake of [ ¹⁸ F]-2 or [ ¹⁸ F]-3 in two TNBC cell lines, (A) MDA-MB-231 and (B) MDA-MB-468. Uptake accumulated over time specifically for epoxidated compound [ ¹⁸ F]-3 and was retained when compounds were removed from the medium 15 min prior to harvest (wash). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Data also presented in Supplementary Table 1.

In Vivo Imaging and Post-PET Biodistribution Studies

We next evaluated the uptake and distribution of the epoxidized [ ¹⁸ F]-3 in vivo in both tumor-bearing and aged matched healthy animals using PET/CT imaging and post-PET biodistribution studies. Athymic nude mice were implanted with MDA-MB-231 or MDA-MB-468 cells, and tumor uptake of [ ¹⁸ F]-3 determined 1 and 4 h post injection after tumors reached a palpable size (Figure A, B). At the conclusion of the in vivo imaging period, animals were euthanized, and organs were assessed for radioactivity. Tumor uptake of [ ¹⁸ F]-3 was the highest in either tumored cohort after 1 h with 0.33 ± 0.11% ID/g in MDA-MB-231 tumors and 0.57 ± 0.22% ID/g in MDA-MB-468 tumors (Figure C).

3.

Axial view PET/CT images of tumor uptake of [ ¹⁸ F]-3 in a representative mouse with an (A) MDA-MB-231 tumor and (B) MDA-MB-468 tumor at 1 h post injection. (C) Tumor uptake (% injected dose per gram) at 1 and 4 h post injection.

Full-body PET/CT imaging of [ ¹⁸ F]-3 in healthy BALB/c mice (n = 3 per time point) as well as in MDA-MB-231 and MDA-MB-468 xenografts (n = 4 per time point) showed the highest uptake in the gallbladder immediately after injection and in the intestines 1 h after injection (Supplemental Figure 4). At 4 h post injection, the tracer continued to clear non-target tissues, as indicated by decreased gall bladder and small intestinal uptake and increased uptake in the large intestine. Biodistribution studies in tissues harvested from untumored control animals 1 h after injection showed the highest uptake of [ ¹⁸ F]-3 in the large intestines at 10.6 ± 0.43% ID/g, which increased to 55.6 ± 14.8% ID/g at 4 h post-injection (Table ). Similar trends were observed in tumored animals with levels in the small intestine of 81.2 ± 10.3% and 38.2 ± 8.29% ID/g in the MDA-MB-231 and MDA-MB-468 models, respectively, 1 h after injection, which accumulated in the large intestine by 4 h with levels of 59.6 ± 14.1% ID/g and 43.0 + 13.4% ID/g (Supplemental Tables 2 and 3).

1. Biodistribution Results .

	1 h post-injection	4 h post-injection
% ID/g	average ± StDev	average ± StDev
blood	0.18 ± 0.02	0.89 ± 1.18
heart	0.23 ± 0.01	0.15 ± 0.07
lungs	0.27 ± 0.05	0.22 ± 0.16
pancreas	0.04 ± 0.01	0.05 ± 0.03
spleen	0.39 ± 0.12	0.5 ± 0.17
stomach	0.57 ± 0.44	0.16 ± 0.10
liver	4.55 ± 2.29	1.46 ± 0.25
kidney	0.47 ± 0.09	0.24 ± 0.05
s. intestine	9.20 ± 9.48	2.94 ± 1.70
l. intestine	10.64 ± 0.43	55.55 ± 14.76
fat	0.12 ± 0.02	0.07 ± 0.07
skin	0.21 ± 0.11	0.66 ± 0.64
muscle	0.90 ± 1.21	0.14 ± 0.09
femur	0.70 ± 0.09	1.27 ± 0.59
brain	0.05 ± 0.03	0.26 ± 0.22

^aData are reported in % injected dose per gram of tissue (% ID/g).

Pharmacokinetic Studies of Taccalonolide AF

In light of the PET imaging and biodistribution data showing high gallbladder accumulation of [ ¹⁸ F]-3, we performed a focused pharmacokinetic biodistribution study of the nonradiolabeled taccalonolide AF, which is modified at C15. The biodistribution of the intact, nonmetabolized taccalonolide AF follows a similar pattern to that of the fluorinated compound, with a rapid loss of free drug in the plasma and liver with a high and sustained accumulation within the gallbladder over 1 h that was accompanied by accumulation in the intestine over time (Figure ). These data demonstrate that the PET signal in animals treated with [ ¹⁸ F]-3 correlates with the accumulation of the taccalonolides in the gallbladder and liver without prior biotransformation or irreversible target engagement. The uptake of both the ¹⁸F radiotracer and the unmodified drug in the gallbladder suggests the highly lipophilic compound is eliminated through the biliary tract without the prior need for any metabolic conversion or irreversible target engagement. We propose that the highly acetylated steroid core of the taccalonolides may be recognized and transported similarly to bile acids and that modifications that inhibit this transport may provide strategies to improve serum and tumor bioavailability.

4.

Plasma and tissue concentrations of taccalonolide AF over a 1 h time-course after 300 ng iv injections to C57BL/6 mice. Data are reported as ng/g tissue (or ng/mL for plasma).

In summary, we have successfully developed a method to radiolabel a mesylated taccalonolide analogue to allow for quantitative uptake and biodistribution studies using a compound that retains the unique biological activities associated with this compound class. This strategy overcomes limitations of prior pharmacokinetic experiments that were complicated by technical limitations of covalent target engagement and analogues with large fluorescent tags previously studied by our group that were not amenable to in vivo imaging. , The greater uptake in MDA-MB-468 tumors as compared to the MDA-MB-231 tumors is consistent with slightly greater uptake of these cells in vitro and a lower IC50 in the former line; however, there is relatively low uptake in both tumor types, suggesting that the primary mechanism to improve tumor delivery would likely be through inhibiting biliary uptake. However, even in light of the relatively low delivery of the radiolabeled compound to the tumor as compared to the gall bladder and intestines, it is important to highlight that systemic delivery of the taccalonolides has been shown to promote antitumor efficacy in multiple tumor models, including in the MDA-MB-231 model that showed the lowest tumor uptake. ,,, This suggests that very little active compound needs to be delivered to the tumor to promote antitumor efficacy and that these compounds may have even greater in vivo potency than we previously appreciated. Our finding that the taccalonolides demonstrate antitumor efficacy in spite of the low levels that reach the tumor with systemic injection combined with studies showing superior efficacy when the drug is directly injected to the local tumor microenvironment suggest that approaches to improve tumor targeting, including modifications that limit biliary uptake and payloads of antibody drug conjugates, could significantly improve the therapeutic window of these compounds to ultimately provide novel effective treatment options for patients with TNBC and other malignancies that have become resistant to currently approved agents.

Safety Statement

No unexpected or unusually high safety hazards were encountered.

Glossary

Abbreviations

[¹⁸F]-2

7-[¹⁸F]­fluorotaccalonolide B

[¹⁸F]-3

7-[¹⁸F]­fluorotaccalonolide AJ

HPLC

high-performance liquid chromatography

DMDO

dimethyl dioxirane

% ID/g

percent injected dose per gram of tissue

SRM

single reaction monitoring

TNBC

triple-negative breast cancer

HER2

human epidermal growth factor receptor 2

ER

estrogen receptor

PR

progesterone receptor.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.5c00208.

• Detailed experimental information, including synthetic procedures; tables containing cellular and tissue uptake of radiolabeled taccalonolides over time; the biochemical activity of the nonepoxidated fluorinated taccalonolide; and representative whole animal PET/CT images (PDF)

∇.

Department of Biochemistry and Molecular Biophysics, Washington University in St. Louis, St. Louis, Missouri 63130, United States

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

This research was supported by R01 CA219948 (A.L.R) and CPRIT RP230360 (D.E.F.). D.E.F. and A.L.R. would like to thank the Max and Minnie Tomerlin Voelcker Fund and the Greehey Endowment for their generous support of our laboratories. Small animal imaging studies were enabled by UAB’s Preclinical Imaging Shared Facility and O’Neal Cancer Center through P30CA013148.

The authors declare the following competing financial interest(s): A.L.R. is listed as an inventor on patents on this class of compounds that is issued to the UT System.